That goal hasn’t changed. But somewhere along the way, we realized the format was holding it back. The go-to resource for a discipline isn’t a chronological feed of articles — it’s a structured body of knowledge you can actually work through. So we’re giving that goal a home built for it.

We’ve started migrating our articles into what we call the Datacop Academy, hosted directly on our website.

What’s actually changing

This isn’t a copy-paste job. Three things are happening to every piece that makes the move:

Articles are becoming lessons. Instead of standalone posts, each piece is being rewritten as a lesson inside a course that covers one specific topic end to end. The articles were always connected in our heads; now they’re connected on the page — each lesson builds on the previous one, and a course reads as one coherent argument rather than a scattered archive.

Everything is being re-examined for the AI era. The last couple of years quietly rearranged a lot of what we do — which parts of the stack carry the weight, how reports get built, how many use-cases a small team can ship. Where an article described the pre-AI version of our work, the lesson now describes the current one. Some pieces needed a light touch; some needed to be rethought from the ground up.

Lessons get room to be interactive. This is the part Substack could never give us. On our own site, a framework doesn’t have to be a static image — it can be something you explore: diagrams you can open up, models you can click through, examples that respond. Where a concept is better shown than described, we now show it.

Where things stand

As of today, 8 lessons are live across two courses, with more in the pipeline. The first course lays out the full map — the technical marketing department, the Pyramid, the stack, the Atlas, and the organizational shape that holds it together. The second goes deep into first-party data: what it is, why it’s suddenly the differentiator, and how to capture it correctly.

What stays the same

The goal, first of all. The Academy is the same ambition — the go-to resource for eCommerce technical marketers — in a shape that can actually carry it: courses instead of a feed, updated instead of archived, explorable instead of static.

And the promise this Substack started with: the knowledge other agencies gate, free.

Some lessons are open to everyone, no questions asked. Others are available on request — free as well, but we ask who you are at the door, because we grant access to people working at eCommerce companies and to our partners. That’s a deliberate choice, and we’re open about the reason: this is knowledge agencies typically keep to themselves, and we’re happy to hand it directly to brands.

You can browse what’s live today here:

👉 datacop.services/academy

We’re back to writing, and we’ll be honest about the gap. The last four months have been some of the most intense of our professional careers, and it’s all down to the one topic on everyone’s mind right now — AI. We don’t think we’ve ever learned or adjusted as much in such a short window. Views on things we’d have called settled moved, in some cases completely.

So there’s a backlog of things we want to share: what we picked up, and where our thinking shifted now that we’re properly inside this AI era.

Today’s is the marketing techstack. And the interesting thing, at least to us, is that the AI era didn’t blow up the stack and hand us a new one. The core components are the same ones we’d have named a year ago. What changed is that a few of them matter more than they ever did — and the way they relate to each other reads differently to us now than it used to.

The stack, in three regions

Here’s how we draw it today.

There’s a live, interactive version at techstack.datacop.cloud if you’d rather click through it, but the shape is simple: three regions, connected by data flowing between them — and, the part we’d have drawn differently a couple of years ago, no single platform sitting in the middle that everything else orbits.

Data Tracking Infrastructure. Its job is to collect clean, complete, compliant first-party data and feed it to the other two regions — which, in the AI era, also means keeping bot traffic out before it pollutes everything downstream. Four functions do the work: server-side collection and routing (MetaRouter), identity resolution and list growth (Retention.com), bot prevention (Human), and email verification (Mailgun).

Data Activation Infrastructure. The hub is Bloomreach Engagement. Its job is delivering personalized, relevant experiences, fast, across every channel — website, mobile app, email, SMS, mobile push, ad audiences.

Data Storage Infrastructure. The hub is the data warehouse — for us, BigQuery. Its biggest source is usually Bloomreach itself: all the customer data you’d expect from a CDP — website behaviour, campaign engagement, offline interactions — flows in here, alongside inventory and orders from Shopify, spend from the ad platforms, and whatever operational data matters. Its job is to hold all of it, cheaply, and to be the layer AI reasons over.

The flow between them is the whole point. Tracking feeds the other two clean data. Storage and Activation then run a loop: the warehouse is the broad, cheap source of truth; Bloomreach is the fast activation muscle; and data moves back and forth between them as needed. There’s no center — three regions, each good at the one thing it exists for.

For years we’d have put one box in the middle and drawn everything around it. What moved us off that is partly the AI era, and it shows up most clearly in the region we’ll start with.

Data & reporting: the region that changed the most

The storage region used to be the quiet one — a cheaper place to park data that didn’t fit in the CDP. It isn’t quiet anymore. For us it’s become one of the most valuable parts of the whole setup, and the reason is straightforward: AI now connects to the data warehouse easily, and once it does, producing quality analytical output gets dramatically easier.

A little context on why the warehouse is the right home for this in the first place. The activation layer runs on a very fast, in-memory database, and that speed costs money — so we keep it lean and don’t hold full data history there. The warehouse is where large volumes of data live cheaply. Bloomreach is usually the single biggest feed into it — all the CDP-grade customer data, website behaviour, campaign engagement, offline interactions — but its scope goes well beyond that: inventory, ad spend, and other operational data all land in the same place. That breadth is exactly what makes AI on top of it so useful — the model can reach across sources that never sat together before.

That unlocks two things, and they’re genuinely different from each other.

The AI analyst, for ad hoc questions. Point Claude at the warehouse and you get something that behaves like an analyst on call. A question goes in as plain language; the model writes the SQL, runs it, and comes back with an answer that reads across customer data, ad spend, inventory, and operational data together. What used to eat the time wasn’t the chart — it was the twenty minutes of “which table, which join, what does this column mean” between a question and an answer. On a well-modeled warehouse, that friction mostly disappears.

This is where documentation stops being a chore and starts being the thing that makes the whole approach work. The AI navigates the warehouse the same way a new analyst would — by reading what’s there. Clear table and column descriptions, a good written overview of how the warehouse is structured and what each dataset means: that’s what lets the model pick the right table and the right join instead of guessing. We used to treat documentation as hygiene you got to eventually. Now it’s closer to the interface — the better the warehouse is described, the better every answer that comes out of it. To illustrate the kind of question this is built to answer — the sort that used to be a ticket to the data team:

“For products where we’ve spent more than $5k on paid in the last 30 days, show me the ones with under two weeks of inventory left at the current sell-through rate.”

To be straight about it, that particular version — joining paid spend, inventory, and sell-through in one breath — is more where this is heading than where most setups sit today. Getting three clean, distinct sources into one warehouse is real work, and it isn’t a given yet. But the shape is the point: one plain-language question spanning data that used to live in three different places, answered without a ticket.

Permanent reports, the kind that traditionally lived in Bloomreach, Tableau, or Power BI. These are the governed, recurring reports the whole team looks at — different from an ad hoc question because the answer needs to be the same every time and trusted over the long run. Building them used to be slow work: model the data, write the SQL, wire up the source, lay out the dashboard. We now build them with Claude Code, which writes and versions the models against the warehouse and helps assemble the report itself. The output is a permanent, reviewable asset rather than a one-off answer.

An example of a recurring report we now build directly in Claude Code with natural language — no Tableau or Bloomreach required: https://5-1-campaign-performance.datacop.cloud/

One thing worth being honest about, because it’s easy to get swept up: AI on a clean, well-modeled warehouse is a genuine step change, but AI on a messy one just produces confident wrong answers faster. The quality of everything in this region is capped by the quality of the data feeding it. Which is a big part of why we now think about the tracking region differently, too — but we’ll come back to that.

Bloomreach: still the activation layer, now with easier assets

Not much has changed in our view of the activation region: Bloomreach Engagement is still the main activation layer, and we still think it’s the strongest option for the job. It’s built around an in-house CDP on that fast, in-memory database, which is what lets personalization resolve in seconds across channels. And because website personalization, email, SMS, push, ad audiences, and the customer database are all parts of one platform, you’re maintaining a single data model instead of six, and you’re not stitching pipelines between separate tools to keep everything in sync.

What did get easier in the AI era is asset creation. The two that used to be the most time-consuming — building email assets and building web layers — have compressed a lot for us, because we’ve built workflows around Claude that generate them and import them straight into Bloomreach. That’s a whole topic on its own, and it deserves its own article rather than a paragraph here, so we’ll go deep on those workflows separately. For now it’s enough to say it’s the reason Claude shows up on the ring around this region in the diagram.

Data tracking infrastructure: quiet, essential, and more important than ever

If the AI era made the storage region more valuable, it made the tracking region more important — because clean input is now the thing standing between you and confidently wrong AI output.

The job of this region is to put complete, clean, compliant, human first-party data into the rest of the stack. We find it clearer to think about it as a set of functions rather than a set of logos — the tool picks are secondary and interchangeable; the functions are what the region has to cover.

Collection and routing. You need behavioural data off the site — cleanly, once, without a dozen tags dragging performance down. That means a single server-side collection point that owns the anonymous ID (so identity lives in your own infrastructure, not in whichever tool fired first, and swapping an activation tool later doesn’t orphan your anonymous users), routes events to every downstream system, and lets you decide where each event goes — high-volume, low-value events like product impressions to the cheap warehouse, the rest to the fast activation layer. The tool we use for this is MetaRouter.

Bot prevention. This is the function the AI era pushed to the front, and the one we’d underplayed before. The web has a fast-growing bot problem, and automated traffic quietly poisons everything: it inflates your numbers, pollutes your behavioural data, and stuffs your lists with signups that were never people. Since something downstream is now reading all of that literally, filtering non-human traffic out at the door has gone from nice-to-have to load-bearing. The tool in our orbit for this is Human.

Identity resolution and list growth. Resolving anonymous visitors to known profiles, stitching the same person across devices and sessions, and growing the addressable list. The tool we use is Retention.com — with the honest caveat we give every client that it’s US-only, because the approach is illegal under GDPR, and they’re open about that.

Email verification. Keeping invalid and junk addresses out of your sends so your sender reputation holds up — which matters more the harder the rest of the region is working to grow the list. The tool we use is Mailgun.

None of this is new in concept. What’s new is the stakes. When a human reads a slightly dirty dataset, they squint and adjust. When AI reads it, it produces a clean, fluent, wrong answer and hands it to you with total confidence. So the least glamorous region on the diagram is the one we’d now argue quietly protects the value of everything else.

Where that leaves us

Same components, rearranged in how we see them. The stack didn’t get replaced by AI; it got a new center of gravity — or rather, it lost its center entirely and became three regions that each earn their place. Storage went from a parking lot to the layer that answers questions. Activation stayed where it was but got faster to build for. And tracking, the quiet base, turned out to matter more than ever now that something downstream is reading the data literally.

That’s how we see it at the moment, anyway. It’s a perspective, not a prescription — this is simply where four unusually intense months landed us, and we wanted to share it while it’s fresh. The interactive model is at techstack.datacop.cloud.

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

What’s the optimal cadence for email campaigns?
Is it 3 campaigns per week? 5? Twice per day?

While we won’t give you a single “magic number,” we will give you a clear framework you can use to find the optimal cadence for your eCommerce brand — including a step-by-step guide you can run yourself.

Situation

Let’s say you want to determine which email cadence drives better results:

• 3 emails per week, or

• 5 emails per week

Step 1: Split Your Subscribers into Two Groups

Randomly assign your subscriber audience into two groups:

• Group A: receives 3 newsletters per week

• Group B: receives 5 newsletters per week

The size of each group is up to you.

For example, if you currently send 3 emails per week and want to test 5, but don’t want to risk major negative effects (such as increased unsubscribes), you can start by sending the higher frequency to only 10% of your list.

In fact, this is the approach we recommend from a risk-minimization standpoint:

• Keep the majority of subscribers unchanged

• Apply the proposed increase only to a small test group

• If the test group shows positive results, scale the change to the full list

Step 2: Start Sending Campaigns Based on the Assigned Cadence

Send campaigns to each group based on the cadence they were assigned.

Step 3: Measure Results

Now it’s time to measure the outcome of your test.

There are two primary metrics you should evaluate:

1. Revenue generated by each group

2. Unsubscribe rate for each group

Let’s assume that after one month, you get results like this:

Step 4: Project the Results Across Your Full List

Next, simulate what results would look like if your entire subscriber base received either:

• 3 emails per week, or

• 5 emails per week

…while keeping the same average revenue per subscriber and unsubscribe rate from your test groups.

Now you can clearly see the tradeoff:

• The 5-email cadence generated an additional $70,000

• But it also resulted in 7,778 additional unsubscribers

One of these outcomes is clearly positive (more revenue), and one is clearly negative (losing more subscribers).

The key question is:

Is the combined effect overall positive or negative?

Step 5: Calculate the Value of a Subscriber

To evaluate the combined impact, you need to calculate the value of a subscriber — since the 5-email group lost more subscribers.

We recently published an article explaining this methodology in detail, which you can find here:

Important: Define the Time Period

When calculating revenue per subscriber, it’s essential to define the time window.

For example, if you compare revenue per subscriber for users who subscribed in November 2024 vs November 2025, the November 2024 cohort will almost always show higher revenue simply because they’ve had more time to purchase.

To ensure a fair, apples-to-apples comparison, you should evaluate subscribers over the same post-subscription timeframe — for example:

✅ Revenue generated in the first 12 months after signup

Example: Revenue per Subscriber (First Year)

Let’s say you want to calculate the average revenue per subscriber within one year of subscribing.

First, select subscribers who joined your list more than one year ago.

Since we’re writing this article in February 2026, we could select subscribers who joined between:

February 2024 – January 2025

Let’s assume you gained 300,000 subscribers during that period.

Now calculate how much revenue those subscribers generated within their first year after subscribing.

Let’s say that number is:

$5,730,000

Now divide revenue by subscriber count:

$5,730,000 / 300,000 = $19.10

So your average revenue per subscriber (within 1 year) is:

✅ $19.10

Step 6: Calculate Incremental Revenue per Subscriber

At this point, you might say:

“Even if these people weren’t subscribed, they still would’ve made some purchases — so we wouldn’t lose the full $19.10 per unsubscribe.”

And you’d be right.

That’s why we need to estimate the incremental value of an email subscriber — meaning the portion of revenue that can realistically be attributed to email marketing.

How to Estimate Incremental Value

A common approach is comparing:

• LTV of customers subscribed to email, vs

• LTV of customers not subscribed to email

A quick (but imperfect) method is dividing customers into two groups and comparing their LTV.

However, subscription status can change over time:

• A customer might subscribe, unsubscribe, resubscribe, etc.

If you want a more accurate report that accounts for this, feel free to reach out — we’d be happy to consult you on how to structure it correctly.

In any case, your analysis will usually result in a table similar to this:

Let’s assume your analysis shows:

Customers subscribed to email have a 24.2% higher LTV than customers who are not subscribed.

So we can estimate that:

• Average revenue per subscriber (1 year): $19.10

• Incremental lift from email: 24.2%

Incremental revenue per subscriber:

$19.10 × 0.242 = $4.62

Meaning:

✅ Each unsubscribe costs you ~$4.62 of incremental revenue over the next year

Step 7: Put It All Together

Now we know:

• Switching from 3 to 5 emails/week generates $70,000 more revenue

• But it also creates 7,778 additional unsubscribers

• Each additional unsubscribe costs $4.62

So the lost value is:

7,778 × $4.62 = $35,934

Net gain:

$70,000 − $35,934 = $34,066

So even after accounting for churn, the higher cadence still produces:

✅ +$34,066 per month in net value

Step 8: Repeat the Process

Now that you’ve learned that 5 emails per week is more profitable than 3, you can repeat the same process to test:

• 5 vs 7 emails/week

• 7 vs 10 emails/week

• different content styles

• different promotional vs editorial mixes

The best part is that this framework applies to almost any email marketing test focused on strategy, not just isolated tweaks within a single campaign (such as testing different subject lines).

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you’d like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a meeting

Suddenly, what had been a “maybe someday” purchase became an easy, immediate yes. The reminder was timely, the offer felt relevant, and the deal made sense.

For e-commerce brands, this is the kind of automation that turns window-shoppers into customers. An open-box automated email campaign doesn’t just fill the inbox with another marketing message - it is a genuine alert for a time-limited opportunity the customer had no other way of knowing about.

At Datacop, we tested this idea with one of our clients - Muziker. Founded in 2003, Muziker’s extensive catalogue is tailored to musicians and content creators of all levels, focusing on supporting creativity and performance through quality gear and expert service. Muziker also operates one of the largest vinyl record catalogs in e-commerce. They deliver to 35 countries and operate 7 physical showrooms in Czechia and Slovakia.

In this article, we will share the results of the automated open box campaign, discuss data and business prerequisites, and explain how the use-case works in Bloomreach Engagement Scenarios.

Results

The evaluation period included more than 40,000 delivered emails from the Open Box campaign, representing approx. 7.8% of all emails delivered during the evaluation. The campaign was compared with a selection of 8 other automated campaigns during the same period.

The performance was the Open Box was unexpectedly strong.

The campaign was the second-best-performing Revenue per Email Delivered email campaign, only after the welcome email. It also had a strong click-through rate of roughly 8.3%. Overall, it was in the Top 3 revenue-generating automated email campaigns during the evaluation period. Although the campaign does not trigger as often as others, such as Abandoned Cart, it engages customers better.

Data Prerequisites

There are a few prerequisites to consider before implementing the Automated Open Box Email Campaign.

First, the Product Catalog needs to be updated with Open Box products as they “arrive”. Ideally, this update occurs daily.

Secondly, each Product_id has to have a way to establish a relationship with the Open Box Item_ids. The method by which the relationship between the IDs is established can differ between projects.

We recommend two options. One way is to feed an extra attribute into the catalog feed. This extra attribute would list the open box IDs of the “unopened” full price version of the product. Another way to establish the relationship is to retain the same product_id as the “unopened” item and a add “-OB” suffix to each id at the end.

Either way, it is important that a relationship between the “open box” version and the full price unopened version of the product can be established

Both methods are shown as examples from product catalogs on the screenshot below

Thirdly, the following business decisions need to be taken before implementing the scenario:

• What does it mean for a customer to “express interest in a product”? - Typically, when a customer views a product at least twice within 7 days or adds it to cart.

• When does a customer “naturally lose” interest in a product? - Usually, this is somewhere between 90-180 days, depending on the e-commerce vertical and type of product.

Implementation

There are three main steps to the logic of the Open Box Automated email.

1. The first step is creating a list of product_ids for each profile, based on their expressed interest. Achieved via jinja and a set attribute node, this step makes sure to assign product_ids that customers express interest in (#34). This list is crucial for the next two steps.

2. The second step is to check daily if any of the products on the “list of products customers expressed interest in” have a corresponding Open Box product that was added to the catalogue! If yes, the scenario generates a trigger event for that customer (#90). The trigger event should contain a list of the Open Box product_ids that are to be displayed in the email itself.

3. The trigger event sends the Open Box Email (#125). Once the email is delivered, remove the IDs in the email from the customer’s list of “a-stock” product_ids.

If you found this post useful…

I hope you found this content helpful and insightful. If you enjoyed it, don’t miss out on future articles and subscribe for free and have our latest content delivered straight into your inbox.

Subscribe now

If you’d like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below.

The 4 most important areas to monitor in email marketing

In our experience, there are four core areas that should be monitored in any email marketing program:

1. Email Newsletter Performance

These dashboards focus on tracking the performance of your newsletter campaigns, helping you understand how your broadcasts engage subscribers and drive results.

2. Email Flows (Triggers) Performance

These dashboards monitor the performance of your automated email flows (or triggers), allowing you to evaluate how effectively each flow supports conversions.

3. Email Collection Performance

These dashboards measure how effective you are at collecting new subscribers across different touchpoints on your website, such as pop-ups, forms, and landing pages.

4. Email Subscriber Performance

These dashboards assess the quality of your email subscribers. They typically answer questions such as:

• How many subscribers convert within a given time period?

• What is the average value of a subscriber?

• How many subscribers remain active and stay subscribed to your newsletter over time?

How we structure email reporting

This framework also reflects how we typically structure our email reporting.

We usually start with an executive dashboard that provides a high-level overview of performance across all four areas. From there, we create dedicated dashboards (or a set of dashboards) that dive deeper into each individual area.

For email flows, we often go a step further by creating a separate dashboard for each flow. This approach provides highly actionable insights and is an invaluable tool for continuously improving the performance of individual flows.

Below is a diagram illustrating how we typically structure email reporting.

Now lets have a look at key metrics that we typically monitor in each of these areas.

Email Newsletter Performance KPIs

When it comes to metrics, we don’t believe there’s a need to reinvent the wheel. At a high level, we focus on six core KPIs that consistently provide the most meaningful insight into newsletter performance.

We typically do not monitor open rates, as this metric has become increasingly unreliable due to a high volume of bot-generated opens.

In addition, we closely track email delivery errors. This includes metrics such as hard bounces and soft bounces, which are critical to keep low—especially for maintaining overall email domain health. These are monitored alongside other system-related error metrics that we aim to keep at a minimum.

Another report we regularly rely on to maintain strong email domain health is a breakdown of newsletter performance by subscriber email health segmentation and subscriber lifetime value (LTV).

This view gives us clear insight into which subscriber segments are responsible for specific portions of the revenue generated.

For example, as shown in the screenshots above, inactive subscribers who have never placed an order received a total of 17 million emails, yet generated only 325 orders.

Based on this, we recommend excluding this segment from email marketing campaigns. Doing so would help improve overall email domain health, which should lead to better inbox placement and, ultimately, higher order volume from newsletter campaigns.

Email Flows (Triggers) Performance KPIs

For email flow metrics, we start with the same core set of KPIs used for newsletter campaigns. However, performance is typically stronger—particularly for metrics such as CTR and revenue per 1,000 delivered emails.

Monitoring errors is arguably even more important for triggered emails, due to the nature of their setup. Unlike one-off sendouts, triggered emails rely on logic defined within Bloomreach scenarios, where errors can occur at various points in the flow.

In this type of reporting, we also monitor daily statistics for each flow to ensure that every flow is triggering as expected. This allows us to quickly identify anomalies and confirm that error rates are not increasing beyond normal levels on a day-to-day basis.

We usually recommend setting up a separate dashboard for each flow. To better understand which metrics are included in these reports, please refer to our article where we take a deeper dive into our framework for evaluating specific triggered flows.

Email Collection Performance KPIs

When it comes to email collection, the most important metric is the number of new subscribers acquired. It’s also valuable to measure this as a proportion of overall new traffic.

In addition, we monitor which email collection touchpoints generate the most subscribers. Based on our experience, checkout is typically the top email collection source for most eCommerce companies.

The report then includes a dedicated section for each email collection touchpoint. For example, this is how we typically monitor an email homepage pop-up banner with two key features:

1. Floating sticky button. This feature allows users to reopen the email sign-up pop-up even after closing it. In the GIF below, it appears as a “Messages” button on the right-hand side of the page.

2. Zero-party data collection prior to email collection. As shown in the example below, we first ask customers about their favorite sport and only then request email marketing consent. This approach often performs better than asking directly for an email address.

The report then highlights several key conversion points:

1. Banner visibility. How many people who saw the “floating pop-up button” actually viewed the banner itself? In our case, this is approximately 75%, a figure that is strongly influenced by reducing the delay before the email pop-up banner opens automatically.

2. First-step engagement. Of the users who saw the first step of the banner (“What’s your go-to sport?”), 13.33% selected an answer.

3. Email marketing consent. Finally, 5.00% of users who viewed the first step of the banner provided email marketing consent.

This report can then be used to support A/B testing of different email collection strategies. We recently published an article outlining the key areas that can typically be A/B tested for homepage pop-up banners and other email collection touchpoints.

Email Subscriber Performance KPIs

Lastly, we have email subscriber performance reporting. We typically track three core metrics:

• Revenue per Subscriber: The average revenue generated per subscriber within a defined time period after subscription.

• Subscriber Conversion Rate: The percentage of subscribers who convert into purchasers within a defined time period after subscribing.

• Unsubscribe Rate: The percentage of subscribers who unsubscribe within a defined time period after subscribing.

It is important to measure all of these metrics within a consistent time window after subscription. Why?

For example, if you compare Revenue per Subscriber for users who subscribed in November 2024 versus November 2025, the November 2024 cohort will almost always show higher revenue simply because those subscribers have had more time to make purchases.

To ensure a fair, apples-to-apples comparison, metrics should be evaluated within the same post-subscription timeframe—such as the first 30 days after signup. This approach allows for meaningful and unbiased comparisons between different subscriber cohorts.

How can you use these metrics?
One of the most valuable use cases is comparing different subscriber cohorts to understand performance differences between them.

For example, you might be evaluating two email collection tactics:

• offering a flat 10% discount, and

• offering a chance to win a weekly giveaway.

Giveaways often generate a higher volume of subscribers, but the key question is whether those subscribers convert and generate revenue in the same way as subscribers acquired through a flat discount.

By analyzing Revenue per Subscriber, Subscriber Conversion Rate, and Unsubscribe Rate, you can answer this question effectively and make data-driven decisions about which acquisition strategy delivers more revenue overall.

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you’d like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a meeting

Additionally, when we conducted an analysis comparing the lifetime value (LTV) of customers who were subscribed to email (and SMS) marketing with those who were not, we observed dramatic differences between the two groups. These results are illustrated in the table below.

If I personally ran an eCommerce store, email collection would be one of my top marketing priorities—primarily for the reasons outlined above. There are additional benefits as well, such as building a strong digital identity, which brings significant long-term value on its own.

In this article, we will explain how we typically approach email collection—starting with placement, and then moving on to how we A/B test different elements of email collection. These elements can have a major impact on the total number of emails collected. We will also share specific client examples along the way.

So, without further ado, let’s dive in.

Email Collection Funnel

The funnel is a well-known concept in the eCommerce space. A classic eCommerce funnel starts with website visitors visiting the home page at the top, then narrows down to users who view a product, add an item to their cart, and ultimately place an order.

Email collection tactics can be mapped onto this eCommerce conversion funnel, which helps visualize all the opportunities to ask a website visitor for their email address.

Different email collection tactics are better suited to different touchpoints within the funnel, although some tactics—such as exit-intent pop-ups—can be effective across multiple stages. Even in these cases, however, you can still leverage the context in which the tactic is triggered. For example, an exit-intent pop-up shown on a product detail page (PDP) can reference the specific product being abandoned (e.g., “Here’s $10 off Product X”), whereas an exit from the homepage typically requires a broader message.

As a visitor moves deeper into the eCommerce conversion funnel, email collection tactics tend to achieve higher conversion rates compared to those at the top of the funnel. Checkout email collection, for instance, can reach conversion rates as high as 60%. That said, the number of visitors at this stage is significantly lower than at the top of the funnel.

The diagram below visualizes this concept and highlights several email collection tactics commonly used at different touchpoints.

Key takeaways from the graph above:

• Checkout is the primary email collection touchpoint for most eCommerce companies, achieving a conversion rate of around 60%.

• The homepage typically has the highest traffic, but even well-performing pop-up banners there usually convert at only around 4%.

• As visitors move deeper into the eCommerce funnel, email collection conversion rates increase. This likely correlates with higher purchase intent and stronger engagement among users who progress further in the funnel.

Optimising Email Collection Tactics

Let’s say you want to increase email sign-ups on your website. How should you approach this challenge?

First, it’s important to understand that there are several distinct elements you can optimise. Broadly, we group them into three core categories.

When it comes to email collection, you can optimise:

• Placement – Where and how you prompt visitors to join your email marketing list across the website.

• Creative – The visual design and messaging of the sign-up prompt itself.

• Offer – The value you provide to a website visitor in exchange for their email address. The key question here is: “Why should I subscribe to your newsletter?”

In the section below, we’ll explain what we mean by each of these categories and share examples from past A/B tests we’ve run within each area.

Placement

Where and how you prompt visitors to subscribe to your email marketing list matters significantly.

Below are two examples from different touchpoints that illustrate the impact of placement and timing on email collection performance.

Example 1: Checkout

A Shopify merchant transitioned from a three-step checkout to a one-step checkout. Importantly, they did not change the wording of the consent checkbox or the offer for subscribing to the email marketing list. The only change was the context in which visitors were asked to subscribe.

As a result, their email subscription rate dropped sharply.

We published a separate article explaining why this happened, which you can find below.

Example 2: Homepage Pop-up Timing (A/B Test)

Another example comes from an A/B test that is currently running for one of our clients. The setup is intentionally simple.

An email pop-up banner is triggered on the homepage. The only difference between the two variants is timing:

• Variant A appears automatically after 3 seconds

• Variant B appears automatically after 5 seconds

Despite the minimal change, the results are clear:

• The 3-second variant collects approximately 15% more email addresses than the 5-second variant.

This demonstrates how even small adjustments in placement and timing can have a meaningful impact on email collection performance.

Offer

The offer defines the value you provide in exchange for a visitor subscribing to your email newsletter.

Common examples of email subscription offers include:

• Exclusive benefits – Early access to sales, sneak peeks at new arrivals, or a consistently valuable and engaging newsletter.

• Flat discount – A straightforward discount that applies to all products or services.

• Giveaway: Offering visitors the chance to win prizes through contests or daily giveaways.

• Choose Your Discount: Let customers select from a range of discount options instead of offering a flat rate.

• Free shipping – Providing free shipping on the first order for new email subscribers.

  

• Wheel of Fortune: A fun, interactive way for customers to spin a digital wheel and win discounts or prizes.

The right offer depends heavily on your brand positioning and business model. For example, if your brand does not offer discounts, a flat discount incentive is not a viable option.

That said, different offers can vary significantly in performance—not just in terms of sign-up volume, but also subscriber quality. In a past A/B test (we can’t share exact numbers), we compared a free shipping pop-up against a giveaway-based pop-up.

The giveaway collected substantially more email subscribers. However, those subscribers proved to be lower quality on average: many users signed up solely for the chance to win, with little long-term intent to engage with the brand.

This highlights an important point: when evaluating email collection tactics, it’s critical to measure subscriber quality, not just the total number of sign-ups.

If you’re interested in how we measure subscriber quality, you can read more about it in this article:

Creative

Creative refers to the wording of your offer and its visual presentation—in other words, how the email subscription prompt looks and communicates value.

We have a few interesting examples that highlight the impact creative changes can have.

The first example involves a simple change to the wording of the email opt-in at checkout. Despite being a minor adjustment, the results were dramatic. You can read more about this case in the article below.

Another example comes from a pop-up banner test. In this case, this creative version:

significantly underperformed this version, despite both promoting the same offer:

Final thoughts

As outlined above, email marketing has a significant impact on the LTV of your customers.
Email marketing strategies begin with email collection.

Optimising email collection is, from our perspective, an ongoing process of A/B testing different aspects of email collection. These can be categorised into three broad areas:

• Placement

• Offer

• Creative

To be successful, this process is typically owned by a dedicated team with clearly defined metrics for measuring email collection success or failure. These usually include:

• Number of email subscribers

• Subscriber conversion rate (how many subscribers made a purchase after subscribing).

• Revenue per subscriber

• Unsubscribe rate (how many subscribers have already unsubscribed)

If you need help with any aspect of email collection, feel free to reach out by booking a meeting through the link below:

Book a meeting

If you found this post valuable…

We hope you found this article valuable. If so, please consider subscribing (for free!) to receive updates on our latest publications.

Subscribe now

This framework focuses on email trigger evaluation. From our experience, when brands assess the performance of triggered email campaigns, they often concentrate on just one area: what happens once someone sees the email. The main question tends to be, “How do I make the email attractive enough for a subscriber to engage?”

While this is an important consideration, it is only one part of the story. There are several other factors—often overlooked—that have a significant impact on the overall performance of triggered email campaigns.

In this article, we will explore these factors by introducing a framework we call the Triggered Performance Funnel.

Triggered Performance Funnel

When you want to improve the performance of your triggered email campaigns, it’s helpful to think in terms of multiple touchpoints—or steps—that significantly influence overall results. These steps can be organized into a funnel.

By increasing the number of subscribers who successfully move from one step to the next at any point in the funnel, you can meaningfully improve the performance of the entire campaign.

Let’s take a closer look at the funnel itself and explore what each step represents, as well as how we can influence the metrics at each stage.

Step 1: From Web Action to Email Sent

Out of all cart abandoners, how many can actually receive a cart abandonment email campaign?

Let’s use a cart abandonment email as an example. In order to send a cart abandonment email, a visitor must first abandon a product in their cart. In theory, this action makes them eligible to receive a triggered email.

However, the critical question is this:
out of 1,000 website visitors who abandon an item in their cart, how many can you actually identify by an email address that has valid consent to receive marketing communications?

Suppose you are able to identify 100 of those 1,000 cart abandoners by email. This means you can retarget only 10% of all cart abandoners. Simply doubling this identification rate would also double the number of emails sent—and, in most cases, double the revenue generated by this campaign.

The key metric to focus on at this stage is the Identification Rate, which can be defined as:

Identification Rate = Number of subscribers entering the cart abandonment flow ÷ Total number of cart abandoners

While this may sound obvious, many brands overlook this step entirely. In practice, we’ve seen identification rates vary dramatically—from as low as 2% to as high as 40%. That difference alone can result in up to a 20× gap in cart abandonment revenue between brands.

The important question now is: what practical steps can you take to increase your identification rate?

This is largely a matter of how robust the digital identity of your website visitors is. We explored this topic in depth in a separate article dedicated to digital identity.

All of the use cases outlined in that article can help strengthen digital identity. However, we would especially highlight cookie lifetime extension, Retention.com, and semi-gated experiences as the most impactful approaches.

Step 2: From Email Sent to Eyeball Connection

Once the email is sent, how likely is it that the subscriber will actually see it?

This stage focuses on how many people who receive your email actually notice it—what we call an eyeball connection.

In simple terms, it’s about increasing the chances that a subscriber becomes aware that your email has landed in their inbox.

There are several levers you can pull to maximize the chances of your email being noticed. The most important factor, however, is inbox placement. If your email ends up in the spam or promotions folder, the likelihood of it being seen is significantly lower than if it lands in the primary inbox.

There are tools specifically designed to improve inbox placement. One such tool is Mailmend.io.

At a minimum, however, we strongly recommend monitoring your domain’s email health using tools such as Google Postmaster Tools.

Another effective strategy is testing different send times and aligning them with the times when individual subscribers are most likely to engage with emails.

Finally, testing different subject lines is highly recommended, as this directly increases the chances of creating an “eyeball connection.”

Measuring the Eyeball Connection

The “eyeball connection” cannot be measured directly, so we rely on a proxy metric. Our assumption is that if a subscriber regularly engages with emails, they are noticing them.

To measure this, we create an Email Health segmentation and calculate the ratio of “New” and “Active” subscribers divided by the total number of subscribers who received the campaign.

Eyeball Connection Rate = (New + Active subscribers) ÷ Total recipients

A healthy benchmark for triggered email campaigns is around 60%.

Step 3: From Eyeball Connection to Email Click

How many subscribers take the desired action after noticing the email?

This is the stage where companies typically focus most of their optimization efforts. That isn’t necessarily a bad thing—as long as it doesn’t come at the expense of the earlier stages we discussed above.

This stage is all about maximizing the chances that, when someone notices your email, they will actually open it and click the call-to-action (CTA) button inside.

The tactics for improving click-through rate (CTR) are among the most well-known and widely used in email marketing. They include subject line testing and improving the overall email copy. “Improving email copy” is a broad concept, and there are many elements you can test, such as:

• Including or excluding product reviews

• Adding a discount or improving the offer for the recipient

• Personalization (for example, reminding a customer to use an unused welcome discount when purchasing a specific item)

• Testing different language, tone, and creative formats

• Adding social proof

…and more.

The primary metric we focus on at this stage is click-through rate (CTR%). A strong benchmark for triggered email campaigns is around 4%.

Brands also commonly track open rate, but in our experience this metric has become increasingly unreliable due to bot-generated opens that occur before the email is actually seen by the user. Since we haven’t found a consistently reliable way to filter out bot activity, we generally do not rely on open rate as a core performance metric.

Triggered Email Performance Optimization Cheat Sheet

For your convenience, we’ve summarized the most important information from this article into a simple cheat sheet. Feel free to grab a copy by downloading the image below.

If you found this post valuable…

We hope you found this article valuable. If so, please consider subscribing (for free!) to receive updates on our latest publications.

Subscribe now

If you’d like to explore any topics mentioned in the article further, or if there’s anything else we can assist you with, feel free to schedule a meeting with us by clicking the button below:

Book a meeting

One underutilized tactic to boost automated email performance is to make use of product tags. These are often displayed on websites in category and product pages, but in emails, they are rarely used!

Recently, we have tested this idea at Datacop with one of our clients - FSA Store. To test, we selected the Cart Abandonment email. Together with their CRM Marketing Team, we added 2 dynamic product tags into the dynamic email template. The tags display in the email above the product if:

• The product left in the cart is low on stock

• The product left in the cart has a lower price than usual

We would like to share the results of the use-case with you and provide further ideas on what kind of product tags you could try out in addition to the 2 mentioned above. Then, if you are still interested, stick around to read more about how it was implemented, the challenges in evaluation, and how we overcame them.

Results

The testing period included more than 100,000 delivered emails from the Cart Abandonment campaign. The email templates contained up to 3 products. In most cases, 81.7%, the products in the email did not contain any tags at all.

The performance of the “tagless” emails provided a baseline to compare the other three variants against. While the number of emails with a price drop product tag was only roughly a fifth of the total, the ones that contained the tag performed better. Combined, emails with at least 1 price drop product tag had a 20% uplift in CTR% and 31% uplift in Revenue per Email Delivered, compared to emails with 0 products tagged.

• Emails containing 1 product with a Price Drop Tag had a 17% and 8% uplift in CTR% and RPE, respectively.

• Emails containing 2 products with a Price Drop Tag had a 27% and 39% uplift in CTR% and RPE, respectively.

• Emails containing 3 products with a Price Drop Tag had a 48% and 296% uplift in CTR% and RPE, respectively.

  

  Unfortunately, during this test, the low stock product tag did not perform very well. Another concern is that this test ran during a period when the catalog had very few low-stock items.

If you come across this situation yourself, we would recommend the following three action points to tackle this issue:

1. Consider whether the threshold used to determine if a product is “low on stock” is set too high.

2. Are there periods of lower and higher inventory counts? Try testing again during a lower inventory period.

3. Would “almost gone” perform better than “Low Stock”? Can a different color of the tag make a difference?

Takeaways & Ideas

The results of this test show the benefit of testing and adding dynamic product tags in your emails.

However, the test also showed that not every tag will be equally useful. Some will perform for you, and some might not. That is why it is important to consider what kind of product tags are relevant for your business. This depends a lot on the type of products you sell and the data available for your product catalog. For instance, if you do not often discount your products, the Price Drop tag may not be relevant for you. Similarly, if you do not have a lot of variation in your inventory, Low Stock tags simply would not display often.

Consider the following Product Tag ideas, which you could test in addition to the 2 tested:

1. Customer Favorite - if an item has a reviewAvg of more than 4.5

2. Just Added / New - if an item's date_added value matches the last 30 days

3. Popular / Trending / Hot - if item’s item_id is within a list of best sellers (Bloomreach Engagement Report Jinja)

4. Limited Edition - if the item is marked as a limited edition product

Implementation

The use-case was implemented in Bloomreach Engagement, thanks to its flexible jinja capabilities within email creatives and its robust data catalogs.

To implement product tags, first confirm that your product catalog contains the necessary data.

For the PriceDrop tag, you need to track two price attributes in your catalog: the original_price and the current_price of the item. Having only the current price is not sufficient.

The Low Stock tag, requires traking and feeding the up-to-date stock_level value per item, ideally in hourly intervals. However, daily updates would be sufficient in some cases.

This catalog data is then referenced in the jinja code responsible for checking if the product that is displayed in the email should also have a relevant tag to be displayed or not.

The second prerequisite to implement this tactic is to have already automated email campaigns implemented, which display products such as:

• Cart / Browse / Category / Checkout Abandonment

• Price Drop / Restock / Low Stock Alerts

• New Arrivals / Best Sellers

The dynamic jinja controlling whether the product tags are displayed or not is added into the Dynamic Block within the selected Automated flows.

For the price drop, the inserted jinja code would loop through the list of item_ids the customer has in their last cart_update event. It would check for each item_id if their a catalog item.original_price > item.current_price.

For the low stock tag, the inserted jinja code would loop through the list of item_ids the customer has in their last cart_update event. It would check for each item_id if their catalog item.stock_level < 20. If the stock level of the item is below the selected threshold value, it would display the tag.

During the implementation phase, it is important to discuss and define how and where the product tag should be displayed within the dynamic block creative.

Evaluation

The evaluation of this use-case is not straightforward.

Due to the dynamic nature of the use-case, out of box, it is impossible to know if a Cart Abandonment email that was delivered actually contained a product tagged with a PriceDrop or LowStock tag, or how many products were tagged.

This presents a challenge to the evaluation - how do we know if the tags are boosting performance?

To answer this, we made use of one of the unique Bloomreach Engagement’s features: custom campaign tracking. Using this feature and a jinja code we added a custom event attribute to all campaign events connected to the emails where the product tags were added, which counted the number of PriceDrop tags and LowStock tags displayed in the email, respectively.

Notice in the screenshot below the campaign event attributes starting with “c_” are custom-added. In this email, from the first screenshot, we can see the data in the screenshot below. Out of the 3 products displayed, 1 had a PriceDrop tag displayed and 1 had a LowStock product tag displayed.

Thanks to these custom attributes, we were able to split out the performance of the Cart Abandonment email with and without the tags, as well as split them out by how many products were tagged and by which tag.

If you found this post valuable…

We hope you found this content worthy of your time. If you enjoyed it, don’t miss out on future articles - subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you’d like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below.

You can watch the full recording here—we hope you find it valuable and actionable!

🚀 What’s Next?

We’re already gearing up for our next webinar, and we’d love to keep you in the loop. If you’d like to be among the first to hear when it goes live, subscribe to our Substack below. All subscribers will get early access and updates as soon as we announce the date.

Subscribe now

One of the central ideas we presented was a categorisation of triggered emails. In this article, we’ll take a deeper dive into that framework. We’ll also provide a comprehensive list of triggered email types that can be incorporated into your email marketing strategy.

Email Marketing Categorisation

Email marketing generally falls into two broad categories: newsletters and triggered emails.

The primary distinction lies in their audience and delivery. Newsletters target a broader audience and are sent on a predetermined schedule, with marketers deciding the specific day and time for delivery.

Triggered emails, in contrast, are sent to specific users based on their recent actions, which serve as the trigger for the campaign. For example, if a user abandons their shopping cart, they may receive a cart abandonment email. These emails typically outperform newsletters because they capitalize on the user’s recent engagement with the brand, making the communication timely and impactful.

Categories of Triggered Emails

Triggered emails can be further categorized based on user actions, events, or data. The following are the main types:

(a) Session Abandonment Flows

Think of these flows as follow-ups that are triggered when someone leaves your website at different touchpoints. From our experience, we’ve identified six types of triggered emails commonly implemented for each of these moments. The key touchpoints include:

(a-1) Home Page Visit

This scenario occurs when someone visits your homepage but doesn’t explore further. Because it reflects the lowest level of engagement on your website, emails triggered at this stage typically show the weakest performance. Still, if you already have the customer’s email address, this moment can serve as a useful trigger for launching a campaign.

(a-2) Search

If your website includes a search function, you can also treat search activity as an abandonment trigger. The advantage of this touchpoint is that you capture the exact search term the customer used during their session. This insight can then be leveraged to personalize the follow-up email with highly relevant content.

(a-3) Collection Page Visit

If a customer visits a specific collection page (for example, men’s t-shirts) but leaves without taking further action, this can also serve as a trigger for an abandonment flow. Since you know which product category the customer was browsing, you can use that insight to tailor the email campaign and make it more relevant.

(a-4) Product Page Visit

When a customer views a specific product but doesn’t complete a purchase, this moment can also serve as a trigger for an abandonment flow. In this type of campaign, you can be highly specific by featuring the exact product the customer was interested in. To strengthen the message, you might also include supporting elements such as customer reviews, ratings, or social proof.

(a-5) Add to Cart

This is perhaps the most well-known abandonment campaign: the cart abandonment flow.

(a-6) Checkout Visit

Some brands choose to separate cart abandonment and checkout abandonment into two distinct triggered campaigns. The key difference is that during checkout, customers often provide additional information—such as a shipping address—that can be leveraged in the follow-up campaign. This extra detail allows for deeper personalization that wouldn’t be possible in a standard cart abandonment flow.

(b) Shop Event Flows

The key difference between session abandonment flows and shop event flows lies in what triggers them. Session abandonment flows are initiated by a customer’s action—such as visiting and leaving the site at different touchpoints. Shop event flows, on the other hand, are triggered by an event within the store itself, like a product coming back in stock.

Take the back-in-stock flow as an example. A customer must first express interest by requesting to be notified when a product becomes available again. However, this action alone doesn’t trigger the campaign—it simply places the customer in the audience pool. The actual trigger occurs later, when the product is restocked, which could be a week, a month, or even longer after the customer’s initial request.

So, what are the typical shop event trigger flows?

(b-1) Back-in-Stock

Imagine a customer browsing your store who discovers that a product they want is currently out of stock. Many eCommerce companies offer the option to be notified once the item is available again. At Datacop, we believe the opportunity goes far beyond this simple use case. By expanding the audience for back-in-stock notifications, we’ve been able to significantly boost performance. In fact, as you’ll see in the article below, we increased back-in-stock revenue by more than 60%.

(b-2) Low-in-Stock

Imagine a customer browsing a product—or even adding it to their cart—that is now running low in stock. This creates a powerful trigger for a campaign that nudges them to complete their purchase before the item sells out.

The key difference from traditional browse or cart abandonment flows lies in timing and context. Let’s say I added a product to my cart but didn’t complete the purchase. I probably received a standard cart abandonment email but still chose not to buy. Now, fast forward two weeks: the same product I showed interest in is suddenly low in stock. At this moment, we send a new low-stock notification email. By combining the customer’s past intent with the urgency of limited availability, we re-engage them at exactly the right time—when they’re more likely to act.

(b-3) Price Drop

A similar idea applies to price drops. Instead of inventory decreasing, the trigger comes from a decrease in price. If a customer previously browsed a product or added it to their cart, and that product is later discounted, this creates the perfect moment to re-engage them. A price-drop email flow leverages both the customer’s prior interest and the new incentive of a lower price—making it a highly effective way to drive conversions.

(b-4)Trending Products

Let’s say a specific product is selling faster than usual. This can serve as a powerful trigger for a campaign targeting customers who previously showed interest in that product, reminding them to act before it sells out.

(c) Customer Journey Flow

Customer Journey flows are flows that are triggered at the significant milestones or events in a customer journey. These are the customer journey flows that we’ve seen in the wild

(c-1) Welcome Flow

The email welcome flow is typically sent to people who have just subscribed to your mailing list. At this stage, it’s important to remember that most subscribers are not yet customers. This makes it crucial to include strategies that lay the groundwork for converting them into buyers.

One of the best opportunities within the welcome flow is collecting zero-party data, which can be used to personalize future communication. For example, here’s how a welcome email might differ when tailored to two different audiences:

(c-2) Post-Purchase Flow

The post-purchase flow is an email sequence sent to a customer after they complete a purchase. It typically consists of multiple emails, each serving a specific purpose. Common objectives include:

• Order Confirmation
  Immediately reassure customers that their order was received with a confirmation email that summarizes purchase details.
  Note: Typically dispatched via an ERP system, though it can also be integrated into Bloomreach Engagement.

• Shipping & Delivery Updates
  Keep customers informed when their order is processed, shipped, and delivered, including tracking details. Timely updates reduce anxiety and prevent support queries.
  Note: Also often sent from an ERP system but can be integrated with Bloomreach Engagement.

• Product Usage Tips & Guides
  After delivery, share helpful content on how to use or care for the product. This improves satisfaction and ensures proper use.
  Note: Importing delivery data into Bloomreach allows you to time these updates perfectly.

• Request for Feedback or Reviews
  Ask for feedback or reviews once customers have had time to use the product. This provides insights for your business while making customers feel valued.

• Cross-Sell & Upsell Recommendations
  Suggest related products or accessories based on purchase history and browsing behavior. Personalized recommendations drive repeat purchases and increase lifetime value.
  Note: Some customers place a second order within minutes of their first. Sending immediate upsell emails—with incentives like free shipping—can capture these opportunities.

• Loyalty & Rewards Invitations
  Encourage customers to join your loyalty program, highlighting benefits such as points, exclusive discounts, or early access to new products.

• Exclusive Offers for Returning Customers
  Reward loyalty with personalized offers or discounts on future purchases to keep customers coming back.

• Replenishment Reminders
  For consumable or time-sensitive items (e.g., skincare, supplements, pet supplies), send reminders when it’s time to reorder.

• Educational Content
  Share how-to guides, industry insights, or trends related to the purchase to keep your brand top of mind and build community.

(c-3) Refund Flow

The refund flow is often overlooked by many eCommerce companies, but it can be a powerful touchpoint for customer communication. When a customer returns a product, it’s an opportunity to:

• Collect feedback: Understand why the product was returned.

• Offer alternatives: If the customer hasn’t placed a subsequent exchange order, suggest alternative products that might better suit their needs.

(c-4) Loyalty Flow

If you have a loyalty program, it’s recommended to create a dedicated email sequence for key milestones within the customer journey, such as:

• Welcome email when a customer first joins the program

• Milestone emails celebrating achievements or unlocking a new tier

• Progress reminders showing what’s needed to reach the next tier and the benefits of doing so

• Expiration reminders for points or benefits that are about to expire

(c-5) Win-Back Flow

This flow targets customers who have purchased from your brand in the past but haven’t engaged or made another purchase for a significant period of time. The main goal is to reawaken their interest—by gathering feedback on why they haven’t returned and encouraging them to start shopping with you again.

(d) Zero-Party Data Flows

These flows are often the most dependent on the industry in which an eCommerce company operates. Zero-party data refers to customer preferences or key information that they willingly share with you. This data can then be used to trigger personalized responses, such as tailored email series.

Here are a few examples we’ve observed across different industries:

(d-1) Birthday Flow

This is an email flow that can be applied across most industries. If you collect a customer’s birthday, you can send them a celebratory email on their special day. Many brands also include a birthday discount code or another type of exclusive offer to make the occasion more rewarding.

(d-furniture) Home Renovation Flow

One of the most valuable pieces of information you can collect as a furniture store is whether a customer is planning—or currently undergoing—a home renovation. If so, it’s a strong signal that they’re prepared to make significant purchases in the near future. Naturally, you want to maximize the chances that they spend the bulk of that budget with your business.

If you gather this information on your website, you can follow up with a tailored email series designed specifically for renovating customers. These emails might include helpful tips for the renovation process, requests for more details about the type of renovation, and personalized product recommendations based on their needs.

(d-baby care) Pregnancy Flow

Imagine you operate an eCommerce platform specializing in baby care products. Knowing the birth date or expected due date of your customer’s baby can transform your business strategy. With this information, you gain a knowledge of the child’s age throughout their entire lifecycle.

This knowledge enables you to personalize the products displayed on your homepage according to the child's developmental stage. It also helps you tailor newsletter content for the upcoming months and informs product selection for retargeting campaigns throughout the customer's journey.

In fact, we’ve seen this approach successfully implemented by a German baby care retailer (though they are not our client). Their email collection banner not only captured standard details but also included a field asking for the baby's birthdate or expected delivery date.

When we subscribed to their email communication, we received this message during the 23rd week of our pregnancy—when our baby was about the size of a banana.

Here’s the email we received during the 40th week of our pregnancy when our baby was approximately the size of a pumpkin.

And this is the email we received when our newborn was 1 month old.

(d-sports wear) Favourite Sport Flow

If you’re a brand that sells sportswear across different sports, collecting information about the activities your customers are involved in can be highly valuable for your email campaigns. For example, if you learn that I practice Brazilian Jiu-Jitsu, you could showcase the specific product line you offer for this sport, recommend related gear such as rash guards and gis used during training, or even provide personalized guidance on how to choose the right gear for my needs.

Email Triggers Overview

We covered quite a few triggered email campaigns today—19 in total. Below is a visual that summarizes all of the triggers we discussed, organized into the four main categories of email triggers we reviewed.

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you'd like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a meeting

To make the most valuable insights accessible to everyone, we’ve created a short, 23-minute video that distills the best parts of the session. So even if you couldn’t attend the live webinar, you can still watch the recording and apply what you learn to your own projects.

🚀 What’s Next?

We’re already gearing up for our next webinar, and we’d love to keep you in the loop. If you’d like to be among the first to hear when it goes live, subscribe to our Substack below. All subscribers will get early access and updates as soon as we announce the date.

Subscribe now

Bella+Duke is the UK’s #1 rated pet food subscription business. Founded in 2017, they create preservative-free recipes tailored to each pet’s individual needs, with a strong focus on improving pet health through natural nutrition.

The use case we’re discussing today began with a fairly common request: to create a new email collection banner for the homepage.

We received a Figma file with the final design from the client and implemented it through Bloomreach Engagement.

There were several elements we really liked about the initial design. Not only were we collecting valuable zero-party data (specifically, whether the customer had a cat or a dog), but we also implemented a “sticky email collection button.” This feature allowed users to easily return to the email pop-up, even if they initially chose to close it.

However, we quickly encountered a problem. Despite these features, the email collection strategy generated a very low number of signups—only about 0.19% of anonymous homepage visitors converted to email subscribers. This result was significantly below expectations. The client asked us to investigate potential issues, and we responded by designing a simple A/B test that was ready within just 15 minutes.

The email pop-up banner was not displayed automatically. Instead, users had to manually click the “Get Free Treats” button located on the left side of the screen to access the email sign-up form.

This led us to a simple hypothesis:
What if we display the pop-up automatically after, say, five seconds? Would that improve conversion rates?

In addition to testing this change in behavior, we also decided to experiment with different copy within the pop-up window. Ultimately, we set up three variants:

• Variant A (Control Group): Users had to click the "Get Free Treats" button to open the pop-up with the original copy.

• Variant B – The email collection pop-up automatically launched after five seconds if the website visitor did not click the “Get Free Treats” button. The pop-up also featured updated copy.

  

• Variant C – Users had to click the “Get Free Treats” button to access the pop-up, which featured updated copy.

To be honest, we expected the email pop-up banner to improve signup rates — but we didn’t anticipate the difference would be this dramatic.

We quickly reached statistical significance with the following result:

Automatically triggering the email pop-up after five seconds (Variant B) generated 41 times more email subscribers compared to the identical version that required users to manually click the "Get Free Treats" button (Variant C).

Variant A (the original version) and Variant C (which differed only in the copy) did not produce significantly different results. At least, we did not reach statistical significance to confidently conclude that a difference existed.

Do Pop-Ups Negatively Affect Conversion Rates?

One common concern with pop-ups is that they might frustrate customers, causing them to leave the website and ultimately harming the overall conversion funnel.

In the context of Bella+Duke, a crucial step toward conversion is the creation of a “Nutrition Plan,” where customers receive a personalized recommendation for their pet food subscription based on their pet’s specific characteristics.

Before reaching the nutrition plan, visitors must complete a nine-step questionnaire that gathers detailed information about their pets.

What surprised us was that visitors who saw the email collection pop-up triggered automatically were actually more likely to complete the full questionnaire compared to those who interacted with Variant A or Variant C.

Although this difference has not yet reached statistical significance, it’s important to highlight that we observed no negative impact on Bella+Duke’s overall conversion funnel — which was the primary concern we set out to monitor.

Continuous A/B Testing

We view A/B testing as an ongoing process for continually improving the eCommerce digital experience. After learning that the automatic pop-up banner delivered significantly better results for Bella+Duke, we were excited to see that the client is eager to continue testing and further optimize the experience.

The next A/B test we will be launching focuses on the creative of the email pop-up banner. Both variants will still trigger automatically after five seconds; however, this time, the visual and messaging of the pop-up will differ.

Which version do you think will win?

Thank You, Bella+Duke Team!

We would like to sincerely thank the Bella+Duke team for their trust and openness in exploring different use cases.

In particular, we want to express our gratitude to Sam Norris, an outstanding CRM Manager. It’s truly a pleasure to collaborate with Sam on brainstorming and testing new ways to personalize the customer experience for Bella+Duke — all made possible through Bloomreach Engagement.

I would also like to thank our Datacop teammate, Samuel Babic, for his outstanding work on this project.

Thank you, and we look forward to launching many more exciting use cases together!

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you'd like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a Meeting

About the Author:
Technical Marketing Consultant specializing in web personalization, email automation, tracking and analytics. Focused on bridging strategy and execution to help marketing teams drive measurable business outcomes. Has worked with brands such as Dedoles, Bella & Duke, and Foodora CZ.

In an era where privacy concerns and cookie restrictions are reshaping digital marketing, businesses must find innovative ways to gather customer insights. One of the most effective solutions? Zero-party data - information that customers willingly share in exchange for a more personalized experience.

With this in mind, we helped a client implement a web-layer solution that focuses on capturing valuable customer inputs and turning them into personalized birthday reminder emails. It’s a practical example of how zero-party data can be translated into meaningful lifecycle communication.

How the Birthday Reminder Web Layer Works

Traditional gifting campaigns often rely on seasonal promotions. But what if customers could proactively tell you who they want to shop for - and when? That opens the space for relevant communication that feels helpful rather than purely transactional.

In this use case, we designed a web layer experience that enables customers to set up birthday reminders for their loved ones.

Here’s how the process unfolds:

1. User Input Collection: The web layer prompts users to enter a recipient’s first name, birthdate, and preferred gift category (e.g., clothing, accessories, footwear).

2. Email Capture (if needed): If the visitor hasn’t provided their email before, they’re given the option to add it for future reminders.

3. Personalized Confirmation: Once the information is submitted, the system confirms the setup and encourages users to add more reminders.

4. Data Storage & Event Creation: Each answer is stored as an event in the customer profile, ensuring future retrieval for email automation.

Smart Automation: Delivering the Right Message at the Right Time

With the data stored in correct format, our automation scenario takes over:

Note: This solution is deployed using Bloomreach Engagement, and the following section explains how to set it up within that platform. While similar results may be achievable with other toolstacks, some limitations may arise if those tools lack equivalent functionality to Bloomreach.

• Triggering the Reminder: A repeating scenario runs daily at a scheduled time, scanning all customer profiles for birthday reminders set 10 days ahead. Customers with a matching event are then passed through the flow for voucher assignment and email delivery.

• Custom Email Personalization: Jinja templates are used to tailor each email with the recipient’s name and birthday, while the chosen gift category drives the recommendations shown - making the message feel personal and contextually relevant. There’s one important consideration, though: each customer can set multiple birthday reminders for different people. Handling this kind of multi-entry logic within a single email requires a more advanced setup. In the next section, we’ll look at how we addressed this challenge using Bloomreach reports.

💡 Pro Tip: Using Reports as a Scalable Workaround for Multi-Entry Personalization in Jinja

The email templates leverage Jinja to dynamically render content for each upcoming birthday. A core challenge in this setup was enabling multiple reminders per customer - a scenario where a single user might create entries for several different people (e.g., a partner, sibling, and colleague).

In Bloomreach Engagement, expression functions like aggregates or running_aggregates are limited in handling nested or repeated data structures. Since each customer could submit multiple reminders - each with a different name, birth date, and category - we needed a more flexible solution.

To address this, we implemented a custom report-based data layer. Each reminder is saved as a dedicated event, and a central report stores these entries using the customer’s email address as the ID key. This setup effectively acts as a quasi-database, allowing us to:

• Maintain multiple reminders per customer

• Store and retrieve each reminder as a discrete object

• Avoid flattening or overwriting data in expressions

This design ensures high fidelity in personalization and makes it possible to iterate over every reminder tied to a given customer during Jinja rendering.

Simplified Jinja Template:

{% set reminders = report_value_by_key('REMINDER_REPORT', customer.email) %}

{% set target_date = (time + 10 * 24 * 60 * 60) | int %}
{% set target_day = target_date | from_timestamp('%d') %}
{% set target_month = target_date | from_timestamp('%m') %} 

{% for birthday_timestamp, data in reminders.items() %}

{% set bday_day = birthday_timestamp | from_timestamp('%d') %}
{% set bday_month = birthday_timestamp | from_timestamp('%m') %}

{% if bday_day == target_day and bday_month == target_month %}
{% for name, preferences in data.items() %}

🎁 {{ name }}’s birthday is on {{ bday_day }}.{{ bday_month }} — here are {{ preferences.category }} gift ideas with your exclusive discount.

{% endfor %}
{% endif %}
{% endfor %}

This logic supports:

• Multiple recipients per customer

• Time-based filtering and targeting

• Dynamic rendering without additional campaign branching

Note: While Bloomreach reports have a known scope limit of 10,000 records visible in the Reports UI, this appears to be primarily a display limitation rather than a strict data cap. In practice, entries beyond this threshold may not be accessible through the report interface, but they can still be referenced in Jinja logic or used in scenarios. While this behavior isn’t formally documented in detail, we've observed that data beyond the visible limit is preserved and remains functional in live use cases — making reports a workable option for storing and accessing larger sets of structured inputs.

Real-World Performance: 2x Higher Engagement Than Top Newsletter Segments

Since its launch, this birthday reminder flow has shown promising results. While it acts more like a personalized service than a traditional marketing campaign, engagement metrics clearly show its impact. The average click-to-delivered rate of this automated flow is around 4 to 7 times higher than the client’s typical newsletter performance.

It’s important to note that the flow only targets users who interacted with the birthday reminder web layer — a naturally more engaged subset of the audience. Still, even when compared to the most active newsletter recipients, the flow performs about twice as well in terms of click-through rate.

Final Thoughts

This birthday reminder use case demonstrates how zero-party data, when captured and activated thoughtfully, can evolve into a scalable lifecycle touchpoint that delivers both customer value and measurable business outcomes. By combining an on-site weblayer with Bloomreach’s automation and personalization features, we built a reminder system that reaches customers at the right time.

While built around birthday reminders, this logic can be applied to other time-based or intent-driven interactions where personalization and timely communication are key.

As privacy regulations and customer expectations continue to shift, brands that invest in collecting and activating zero-party data will be best positioned to deliver personalization that feels timely, respectful, and useful.

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you'd like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a meeting

About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

The search landscape is undergoing a fundamental shift. Search engines like Google and Bing have established themselves over the past 20 years as the gateway to the rest of the internet. Search Engine Optimisation (SEO) became a core competency of digital marketing teams in ecommerce companies. However, SEO, Google traffic and generally customer behaviour when “searching information online” is likely to experience major changes in the foreseeable future.

Why?

The advent of generative AI (GenAI) search tools such as ChatGPT, Perplexity and DeepSeek are emerging as new channels for consumers to discover products and services online. A few months after their release, these platforms added the functionality to connect to “current up to date internet”.

For customers who are shopping online, this functionality enabled searching the web with more complex and specific queries compared to tradidional keyword-based searches. This is offering a new method for online customers to find what they are looking for - with more refined and intent-browsing experience.

This is opening new opportunities and challenges for e-commerce companies and their digital marketing teams. As a result CMOs are starting to ask themselves and their teams the following critical questions:

• Are users using Search GenAI platforms to discover e-commerce products?

• Are users finding our products via Search GenAI platforms?

• If these users come to our site from a Search GenAI platform, do they convert?

• How do we make sure our website is listed in GenAI Search results?

• Are there UX implications on the site for users who discover the shop from a GenAI search tool?

• What implications does this have for CMOs and e-commerce marketing team?

In the article below we’re going to outline 4 activities CMOs and e-commerce marketing teams can do answer the questions above and prepare for the changes ahead.

1. Identify and Monitor GenAI Search Traffic

Recently, GenAI search platforms are starting to tag their external links with UTM tags so they are becoming possible to measure.

The first step to prepare is to identify and monitor GenAI search traffic on your e-shop. This requires the setting up new traffic groups in Traffic Sources definitions. In Bloomreach this is done via an Event Segmentation.

With the help of the above traffic definitons build a dashboard that allows you to see:

• Count of GenAI Search Session_starts over time (per week / per month)

• What type of pages are they landing on? (Homepages, PDP, PLP, etc.) Which of our categories are receiving traffic and which are not?

• What % of those sessions arrived for the 1st time to the site?

• Conversion funnel - what % of GenAI Search ends up purchasing? If so, which products?

At Datacop we are monitoring GenAI traffic for our clients and we share the following insights:

• Traffic from these platforms still represents less than < 1 % in all cases, however it is growing fast.

• Traffic from these platforms are converting at a similar or better rate as the average of the “normal” traffic.

• Customers who visit via Perplexity, while being a relatively minor platform; tend to have a higher rate of conversion to purchase than average traffic sources

2. E-commerce: Preparing to Adapt

As GenAI search tools become more prevalent, e-commerce teams need to proactively ensure their content is discoverable. Their adoption is growing quickly. It took ChatGPT just 5 days to reach 1 million users. For comparison, it took Instagram 2.5 months to reach 1 million users. By January 2023, just 3 months after the launch, ChatGPT counted 100 million users. For comparison, it took TikTok 9 months to reach 100 million users and it took Instagram 2.5 years to reach 100 million. By February 2025, OpenAI has reported over 400 million weekly active users of ChatGPT (reference).

With such strong growth and proliferation of GenAI-Powered solutions, it is not hard to foresee that while GenAI Search traffic in e-commerce is still more of a curiosity, it could soon represent a considerable % of “Organic Search” of traditional search engines such as Google, Bing or Yahoo. Their growth is underpinned by the fact that it offers truly new and revolutionary ways of interacting with the vast amount of information available to “surf” on the web. Additionally, when consumers were surveyed for which use-cases consumers would consider using search engines with “AI-generated results”, the most common answer was “Online Shopping”.

In how many years will this traffic become a significant portion of your website’s traffic? In one, two maybe three years? The answer to this questions, is the amount of time your team has to adapt to these changes.

ChatGPT and the like offers online shoppers a great way to find specific products on the wide web quickly. These are more likely to be considered purchases, rather than impulsive ones; as the shopper needs to have an intent in mind to find your site via these platforms. As a result At Datacop, we believe that e-commerce verticals that have 1) high AOV 2) high product specificity 3) lots of choices available on the market are at high risk of disruption caused by GenAI Search platforms.

CMOs of these verticals should start considering the following:

• Who on my team is responsible for GenAI SEO?

• What needs to be done so that our site is listed in GenAI Search platform results?

• What prompts do we want to “be listed for”?

• Are there any UX implications for our own search capabilities?

• What does this mean for our web / product content and structure?

3. SEO for your e-commerce pages

As GenAI search tools become a growing source of online discovery, optimizing your e-commerce pages for AI-driven search is essential. While traditional SEO principles still apply, adapting to the way AI platforms process and retrieve information will be key to maintaining visibility and relevance.

1. Use Comprehensive Schema Markup – Mark up every product with details like name, price, availability, description, brand, SKU, and especially aggregate ratings and reviews. AI engines depend on this structured data for precise answers​. For example, if a user asks “What’s the price of Acme Smartwatch X?”, a search AI can confidently answer if it sees structured data with your price. Include Review schema to feed the AI with average rating and review count – generative answers often highlight “Rated 4.5 stars based on 200 reviews” if they have that info. (reference)

2. Write Descriptive, In-Depth Product Descriptions – Unlike keyword-focused SEO, GenAI tools prioritize content that provides context and detail. Ensure product descriptions are rich, informative, and written in natural language to align with how AI models process information. Including detailed use cases, comparisons, and specifications can improve visibility. (reference)

3. Leverage Customer Reviews and Ratings – Showcase customer review content on the page (and not just in a carousel that’s hidden to crawlers). For instance, highlight a “top review” that discusses the product’s use. AI models integrate data from reviews when forming answers​. A question like “Is the battery life good on this smartwatch?” could be answered by the AI quoting a common sentiment from reviews (“Many users report the battery lasts 2 days on a single charge”). By having reviews text accessible, you increase the chance the AI extracts positive attributes of your product from real users. (reference)

4. Incorporate FAQs or Q&A Sections – Many product pages now include an FAQ section (“Q: Is this compatible with iPhone? A: Yes, via Bluetooth.”). This is excellent for AI SEO. It provides clear question-and-answer pairs that an AI can directly grab for voice search or chat responses. Consider adding a small FAQ under each product with 3-5 common questions about the item (compatibility, materials, warranty, etc.) and concise answers. Mark these up with FAQPage schema for bonus points. This structured Q&A content makes the AI’s job easy when a user asks a similar question. (reference)

As search behavior evolves, e-commerce brands that proactively adapt their SEO strategies for AI-powered discovery will gain a competitive edge. By tracking GenAI search traffic, engaging with AI-referred visitors, and optimizing content for AI-driven indexing, CMOs and digital marketing teams can position their brands for success in this rapidly changing landscape.

4. Ask Visitors for Their AI Prompts

If your techstack contains Bloomreach Engagement, you can do more than just identify and monitor the traffic on your site and beginning to prepare a plan for GenAI SEO for your site.

With Bloomreach Engagement’s website targetting capabilities, we recommend to target users that have started their session via a link from ChatGPT and others with a weblayer asking them to share the prompts used that led them to the site.

Understanding what prompts are leading the GenAI platforms to link to your site is valuable information to better understand how users are discovering your site and what their shopping intent is. If you find that users are not sharing the prompts in large enough quantities, consider A/B testing a discount offering to incentivise users to sharing their prompts at higher rates. The prompts will be valuable information that could allow you to have a competitive edge over your competitors for visibility in results on these platforms. They will provide valuable feedback on the GenAI SEO needs of your site:

• What kind of Questions do we need to answer on our product pages?

• What information is important to contain within the product descriptions?

Additionally, if you have a customer services team, you could send the prompt responses with the customer’s contact information via a Scenario webhook to further assist these potential high value customers in their shopping journey via a phone call.

Similar to how we recommend to set up the Scenario webhook alerts in this article below, except for B2B customers, target users who have a session_start which arrived from one of the GenAI Search platforms.

If you found this article useful

If you’d like to discuss the changes in GenAI Search and what you can do to prepare with Bloomreach with us or any other questions related, feel free to book a free 30 minute meeting at the link below.

Book a Meeting

If you enjoyed it and don’t want to miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

Dan and I (co-founders of Datacop) actually used to work for a company called Exponea, which was later acquired by Bloomreach. The platform Exponea developed was then rebranded as Bloomreach Engagement.

We left Exponea about a year before the Bloomreach acquisition to start our own agency. At the time, our relationship with Exponea wasn’t exactly at its peak, so naturally, we explored alternatives on the open market.

To our surprise, we couldn’t find anything that even came close to what we were used to with Exponea (now Bloomreach Engagement). That’s ultimately why we decided to stick with the technology. Fortunately, over the years, our relationship with Exponea (now Bloomreach) has improved significantly.

Of course, you might argue that since we worked for the company that built Bloomreach Engagement, we’re biased—and you may not be wrong. But at Datacop, we prioritize transparency in everything we do, and this article is no exception. We’ll openly lay out all the reasons why we believe Bloomreach Engagement is superior to other alternatives. In the end, you can judge for yourself whether our argument makes sense or if you find it overly biased.

Our reasoning for recommending Bloomreach Engagement over other tech stack alternatives has evolved over the years—and it will likely continue to evolve. This article presents the latest iteration of our argument for why we believe Bloomreach Engagement stands out among its competitors.

Hope you enjoy the read!

What Is Bloomreach Engagement?

Bloomreach Engagement is a unique platform that consolidates the functionality of multiple standalone tools into a single, tightly integrated system. In our view, the platform currently covers ten core functionalities, which we’ve visualized in the following diagram.

We believe there are several major advantages to this all-in-one approach compared to purchasing multiple separate tools and integrating them manually:

1. Reduced Integration Complexity

With Bloomreach Engagement, you only need to integrate one tool into your existing setup. In contrast, if you opt for five standalone tools to match its functionality, you must integrate all five with your infrastructure—and with each other. Maintaining these integrations so they work as expected at all times adds ongoing complexity.

There’s also a significant upfront cost to setting up these integrations, along with recurring expenses for maintenance. From our own experience, we worked with a client who transitioned from Bloomreach Engagement to a collection of separate tools. After the switch, we noticed a clear shift in the type of work we were doing for them—we spent far more time maintaining data pipelines between multiple systems and much less time on revenue-generating use cases and strategic initiatives.

2. Simplicity

Instead of juggling five different logins and interfaces across multiple platforms, you only need one.

3. Cost Efficiency

From our experience, purchasing multiple platforms to approximate Bloomreach Engagement’s functionality is typically more expensive than simply paying for Bloomreach Engagement itself. Additionally, maintaining Bloomreach Engagement and deploying new use cases within the platform is likely 40%+ cheaper compared to managing an alternative tech stack. This cost efficiency is a major factor that businesses should consider.

4. Flexibility

Bloomreach Engagement is incredibly flexible. If you have a use case in mind, chances are you can execute it within the platform. Unfortunately, we’ve found that this level of adaptability is often lacking in other alternatives we’ve tested.

Are There Any Disadvantages?

Because Bloomreach Engagement offers unparalleled flexibility and consolidates at least ten core functionalities into a single platform, it is naturally more complex than a point solution like Klaviyo. However, this challenge can be easily addressed by hiring the right skill set within your technical marketing team—something we’ve covered in this article:

When implemented successfully, this approach significantly enhances your ability to execute revenue-generating use cases. In fact, compared to alternative tech stacks, the impact on your business could be an order of magnitude greater.

10 Core Functionalities of Bloomreach Engagement

In this section, we’ll discuss the architecture and key functionalities of Bloomreach Engagement.

CDP (Single Customer Profile)

One of the smartest decisions Bloomreach Engagement made early on was structuring its platform around a true Customer Data Platform (CDP) at its core. From there, they built out-of-the-box marketing channels such as email and website CRO & personalization around this CDP.

I emphasize “true” CDP because many companies claim to have one, but in our opinion, their products don’t fully meet the definition. We explain this argument in detail in this article:

It’s also worth noting that some companies started as single or multiple-channel marketing tools (e.g., Braze initially focused on mobile apps) and later expanded to support additional channels. Others, like Klaviyo, attempted to bolt on a “CDP” after the fact. However, it appears that adding a CDP as an afterthought is far more difficult than designing the platform around a CDP from the very beginning and then building marketing channels on top of it.

In many ways, Bloomreach Engagement’s CDP-centric architecture directly influences the functionality of all its marketing channels. Because the CDP was designed properly from the start, we haven’t encountered any significant limitations—whether in the individual marketing channels or in how they work together within the platform.

The Power of a Single Customer Profile in a CDP

One of the core features of a Customer Data Platform (CDP) is the creation of a single customer profile—essentially a centralized record that consolidates all important customer or website visitor interactions. This profile aggregates data from various sources, whether it’s offline purchasing behavior, online browsing and shopping activity, or interactions with your email and SMS marketing campaigns.

Since email and SMS are natively integrated into Bloomreach Engagement, there’s no need to manually sync data from these channels into the CDP—it’s already there out of the box. Beyond that, the profile can also include insights such as which A/B test group the customer was part of, what products were recommended to them on your website, and more.

Here’s an illustrative example of what a single customer profile looks like within Bloomreach Engagement:

The Challenge of Maintaining a Unified Profile

However, it’s important to acknowledge that maintaining a single, accurate customer profile in today’s digital landscape is not an easy task. This isn’t a limitation of tools like Bloomreach Engagement but rather a consequence of how customers interact with websites across multiple browsers and devices.

There are many strategies to maximize the accuracy of single customer profiles, which we’ve outlined in detail in this article:

When it comes to executing revenue-generating use cases or answering critical business questions through data analytics, the effectiveness of these efforts heavily depends on the quality of the data they rely on—as illustrated in the visualization below.

This means that both the quality and quantity of data in your database directly determine your ability to drive revenue and make well-informed decisions. That’s why having a strong, high-quality database at the center of all your marketing operations is absolutely crucial.

In the case of Bloomreach Engagement, this central database functions as a highly flexible and incredibly fast CDP, ensuring that every connected marketing channel operates with accurate, unified, and actionable data—ready to be leveraged in real time.

This means that if a website visitor performs a certain action, you can respond almost immediately within the same session. This level of real-time responsiveness is not a standard feature across all marketing operations tools available on the market today, making it a significant advantage of Bloomreach Engagement.

Zero-Party Data

In addition to tracking customer or website visitor interactions across your site and marketing channels, Bloomreach Engagement allows you to store zero-party data (ZPD) within the same customer profile.

Zero-party data represents explicit customer preferences that are voluntarily shared—typically through direct questions such as:
"What is your shoe size?"

Bloomreach Engagement provides a comprehensive infrastructure for collecting ZPD across multiple channels, including your website, email, SMS, and more.

We’ve actually written a detailed guide on collecting ZPD, which you can find in this article:

It also includes free templates we’ve developed, which you can use in your Bloomreach Engagement project—free of charge:

Email Marketing

Everything you’d expect from an email marketing automation tool like Klaviyo is also available in Bloomreach Engagement.

In fact, we believe that email marketing executed through Bloomreach Engagement outperforms Klaviyo—and we explain our reasoning in detail in this article:

SMS (WhatsApp) Marketing

Bloomreach Engagement includes a robust SMS and WhatsApp marketing channel, eliminating the need for third-party integrations. Traditionally, businesses would have to rely on tools like Attentive to access this functionality, but with Bloomreach Engagement, it’s natively built in—allowing for seamless execution and data unification across all marketing channels.

Mobile Push Notification

If you have a mobile app where customers can shop outside of your website, Bloomreach Engagement has you covered. The platform enables you to send personalized push notifications directly to your customers, keeping them engaged and driving conversions.

Paid Retargeting

Bloomreach Engagement also allows you to seamlessly sync audiences built within your CDP to Facebook, Google, and other advertising platforms. This enables you to retarget customers based on first-party & zero-party data that advertising platforms typically don’t have access to.

Additionally, you can exclude customers already subscribed to email, SMS, or mobile push notifications—helping you reduce ad spend and improve efficiency.

Data Analytics

Bloomreach Engagement features a powerful analytics module that provides exceptionally fast data processing. Thanks to its real-time data storage, reports and computations are incredibly fast, making it an invaluable tool for ad-hoc analysis.

Unlike many competitors that operate on a 5- to 15-minute refresh window, Bloomreach Engagement provides true real-time reporting, allowing you to make data-driven decisions instantly.

Mobile App CRO & Personalization

Beyond push notifications, Bloomreach Engagement also enables real-time personalization within mobile apps, allowing you to dynamically tailor the user experience.

For example, if a customer adds a product to their cart, you can dynamically update their hero image based on the product they’re interested in. We originally developed this use case for website personalization, but the same approach can be applied to mobile apps as well. You can read more about it in this article:

Web CRO & Personalization

Bloomreach Engagement is an exceptionally powerful tool for website CRO (Conversion Rate Optimization) and personalization. It allows you to implement a wide range of use cases, from simple customizations—such as displaying a welcome discount to new visitors across the website and other channels—

— to more advanced strategies, such as delivering entirely different homepage experiences tailored to specific customer cohorts. With Bloomreach Engagement, you have the flexibility and infrastructure to execute both quick wins and ambitious personalization initiatives at scale.

Notification System

As proof of Bloomreach Engagement’s flexibility, we’d love to showcase a custom functionality that we at Datacop have built on top of the platform.

We developed a comprehensive revenue loss prevention notification system with two main objectives:

1. Monitoring Marketing Automations & Data Flows – Ensuring that all workflows are running as expected and catching any potential issues early.

2. Detecting Anomalies – Identifying both positive and negative trends in your data that are worth investigating.

We’ve described this solution in detail in the article below, where we break down how it works and how it helps businesses stay proactive in preventing revenue loss:

Anything Missing?

If there’s a channel or functionality not natively available in Bloomreach Engagement, you can easily integrate third-party tools to fill the gap.

Additionally, Bloomreach Engagement offers several other features that we didn’t cover in dedicated sections, such as deploying a simple Loyalty Program and Data Predictions. While these functionalities may not be as widely used, they are still available within the platform.

As we mentioned at the beginning, Bloomreach Engagement is an incredibly flexible platform. If you have a specific use case in mind, chances are you can execute it within Bloomreach Engagement—a statement we probably wouldn’t make about any other tool on the market.

If you found this post valuable…

We hope this article was valuable to you. If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

If you'd like to discuss any aspect of eCommerce technical marketing with us or explore potential collaboration, feel free to schedule a meeting using the link below:

Book a meeting

About the Author:
Co-founder of Datacop, agency that fulfils technical marketing roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

When someone asks me what a great shopping experience looks like, I always point to the brick-and-mortar IKEA shopping experience as a prime example.

I don’t know about you, but I know many people who genuinely enjoy shopping at IKEA stores. The way they've designed their customer journey is truly remarkable.

However, I believe the IKEA in-store shopping experience has one major limitation: physical space.

It doesn’t matter whether a customer owns a house with a garden and pool in the suburbs or lives in a two-bedroom apartment in the city center—the IKEA store experience is virtually the same for both.

This is because there's only one physical version of an IKEA store, making it a static experience for every customer cohort.

eCommerce, on the other hand, offers a unique advantage. Without the constraints of a physical space, we can tailor distinct shopping experiences for different customer cohorts in the digital realm. A first-time visitor to our store might see a completely different homepage than a returning customer who has placed multiple orders. To learn how we design these personalized homepage experiences, read the article below.

A digital IKEA shopping experience can be tailored to a customer's living situation—for example, someone with a house, garden, and pool might see more outdoor furniture, while a person in a two-bedroom city apartment would be shown space-saving solutions.

Creating these personalized experiences across multiple channels—like email, SMS, or websites—for different customer segments in the digital space largely falls under the responsibility of a technical marketing department, which is our area of expertise.

Companies typically focus on personalized shopping experiences because it is one of the key strategies for maximizing customer retention, a responsibility that again often falls under the technical marketing department.

In this article, we’ll dive into the technical marketing department: what its main areas of expertise are, who typically works in this department, and the technologies they rely on.

So without further ado, let’s get started!

Technical Marketing Department

We’ll start by outlining the key areas of expertise that define a strong technical marketing department. These areas can be structured into a three-tiered pyramid, each level representing a different aspect of expertise. Here’s what it looks like:

Data Collection & QA

Think about it—if you want to personalize the customer shopping experience, you first need to know something about the customer. This information is typically stored as data.

In eCommerce, several types of data are particularly valuable:

• First-Party Data – This comes from observing how customers interact with our website and engage with our marketing activities.

• Zero-Party Data – This is data customers share proactively, such as their preferences: “What is your shoe size?” “Which colors do you prefer?” or “What is your living situation?”

• Contact Information – Typically an email or phone number, allowing us to reach customers through marketing efforts.

This foundational data is the bedrock of all personalization efforts, which is why Data Collection & QA sits at the base of the pyramid. The more robust your database, the more opportunities you have to understand your customers and execute effective personalization strategies.

We've decided to share much of our knowledge on these topics for free through our Substack, in the form of articles—just like the one you're reading now. We've already published several articles covering best practices related to the foundation of the pyramid, which you can find below. We hope you find them valuable!

First-Party Data Collection & Data Quality:

Zero-Party Data Collection:

Email & SMS Collection:

Data Analytics

Next, we have Data Analytics. We view data analytics as having similar utility and characteristics to vision. Just as vision helps you navigate the world, strong data analytics frameworks and tools enable you to clearly understand what’s happening in your business and identify the factors driving positive or negative outcomes.

We don’t view data analytics as binary—something you either have or don’t have. Instead, the quality of data analytics exists on a continuum.

It's much like vision. The world isn’t divided into those who see perfectly and those who are completely blind. Most people fall somewhere between these extremes. For instance, you might have mild nearsightedness, allowing you to see and understand your surroundings well, but with occasional blurriness. The same concept applies to data analytics. Any eCommerce company exists somewhere on this continuum, ranging from being nearly "blind" to having a crystal-clear understanding of the factors driving positive and negative outcomes in their business.

Of course, we don’t claim to have created the perfect analytical tools or frameworks (which may not even be possible) that can answer every possible question.

Instead, we offer the best approximation of eCommerce reality that we’ve developed up to this point.

To learn more about our approach to data analytics, we recommend continuing with this article:

Data Activation: Revenue Generating Use-Cases

Finally, at the top of our pyramid, we have revenue-generating use cases. These can take many forms—for example:

• Marketing Automation Campaigns – These can be either personalized or generic for all customers. (For instance, you can find an overview of automated email flows in this article.)
  

• On-Site Personalization – Personalization tactics can be deployed at specific touchpoints, such as the website. (Example below.)
  

• Unified Omnichannel Experience – Ensuring a seamless, personalized experience across multiple channels. (See the use case example below.)
  

These are just a few examples of how businesses can leverage data to drive revenue.

If you’d like to learn which revenue-generating use cases we’d recommend for your specific context, feel free to book a complimentary meeting here:

Book a meeting

Technical Marketing Roles

The Technical Marketing Department typically sits at the intersection of the Marketing and IT Departments.

To bridge the gap between marketing strategies and the technical infrastructure that supports them, the technical marketing team must have a strong understanding of both areas. The diagram below illustrates key functions of the Marketing, Technical Marketing, and IT Departments based on our experience.

We believe there are six key Technical Marketing roles, which we also offer as a fractional, on-demand Technical Marketing department for our clients. These roles include:

Below is a detailed explanation of each role:

Website Personalization & CRO Specialist:

This role is ultimately responsible for optimizing the website experience for its visitors. The primary metric they aim to maximize is average revenue (or profit) per website visitor.

Note: Revenue per website visitor is a more insightful metric than conversion rate. For example, if you raise the prices of your products, the conversion rate might decrease, but revenue (or profit) per visitor could increase. This metric provides a clearer indication of whether the price increase was a good decision.

This individual should continuously consider how the website experience should be tailored for different website visitor cohorts. Every modification to the website should be A/B tested. If the test demonstrates a positive impact on visitor behavior, the change should be implemented; if not, it should be rolled back.

This role typically collaborates with a more technical team member (in our case, a Martech Engineer) who handles the coding for more complex website personalization use cases.

Owned Channels Specialist:

The Owned Channels Specialist is responsible for managing all automations across owned channels, including cart abandonment workflows, back-in-stock and low-in-stock notifications, and other automated campaigns. This role also oversees email collection strategies and personalization initiatives to optimize engagement and retention.

Key performance metrics for this role include:

• Size of the active subscriber base for owned channels

• Revenue per subscriber for owned channels

Additionally, this specialist leads the newsletter marketing strategy, ensuring that content is effectively personalized for subscribers. They are responsible for implementing personalization tactics within newsletter sendouts and developing strategies to collect subscriber preferences through owned channels.

Business Intelligence (BI) Specialist:

The Business Intelligence (BI) Specialist is responsible for the Data Analytics layer of the pyramid depicted earlier in the article. As mentioned earlier, we view data analytics as akin to vision—just as vision helps you navigate the world, a strong data analytics framework (and specialists) enables you to clearly understand your business and identify the factors driving positive or negative outcomes.

This specialist is the go-to person for analyzing business performance, uncovering key drivers behind success or underperformance, and translating data into actionable insights.

The BI Specialist plays a critical role in supporting:

• Owned Channels Specialists

• Website Personalization & CRO Specialists

• Other departments, such as Performance Marketing, by providing the data-driven insights necessary for informed decision-making.

CRM Consultant:

A CRM (Customer Relationship Management) Consultant is someone who takes a strategic, channel-agnostic approach to managing customer relationships. Their core responsibility is to develop a cohesive retention strategy that spans multiple touchpoints, including email, SMS, website experiences, and even paid retargeting. They work closely with Owned Channels Specialists, Website Personalization & CRO experts, and Paid Retargeting teams to ensure alignment. The consultant's primary focus is to make sure personalization strategies are consistent and integrated across all relevant channels.

Data Engineer:

This individual is ultimately responsible for the quality and scope of the data collected by an eCommerce company. They define the data structure that needs to be captured, typically in the form of a tracking document (e.g., view_item and view_item_variant tracking documents), and ensure that the data is consistently and accurately collected. To achieve this, they oversee data quality processes, such as maintaining Slack notifications for monitoring data integrity.

MarTech Engineer:

The Martech Engineer is responsible for integrating the marketing tech stack, including Customer Data Platforms (CDPs), email and SMS automation platforms, website personalization tools, and other marketing automation systems. This role also involves creating and maintaining data warehouses and pipelines that connect these systems.

As the most technically demanding role, it typically requires strong proficiency in SQL and Python for managing data infrastructure and automation.

Beyond integration, the Martech Engineer plays a critical role in implementing advanced personalization use cases, often handling the necessary coding for these integrations. For this, expertise in JavaScript, HTML, CSS, and a strong understanding of web technologies is essential.

Backend vs. Frontend Martech Engineers

In practice, we typically have multiple Martech Engineers with different skill sets:

• Backend Martech Engineers specialize in marketing tech stack integration and data warehouse management, focusing on data pipelines, automation, and system connectivity.

• Frontend Martech Engineers specialize in website personalization use cases, requiring strong JavaScript, HTML, and CSS skills to optimize user experiences and dynamic content delivery.

Technical Marketing Roles - Final Thoughts

We believe that the technical marketing team should function as a cohesive unit, closely collaborating across all areas. From our experience with eCommerce clients, we’ve often seen separate teams handling different aspects of technical marketing with minimal communication: one team manages owned channels, another handles website experience, and a third oversees business intelligence. This siloed approach can lead to friction. For example, the owned channels team might want to include a 'subscribe to newsletter' checkbox at checkout, while the website team opposes it to maintain a clean user experience. Sometimes these teams focus more on optimizing their specific KPIs rather than considering what’s best for the overall customer experience.

We believe that this separation can result in an inconsistent customer experience across different channels, which is far from optimal. Since there is only one customer experience delivered across multiple channels, the teams should have a unified objective and work closely together to achieve it.

This is one of the reasons we favor Bloomreach Engagement as a tool. It integrates six different tools into a single platform, ensuring that all technical marketing roles work within the same system, which naturally fosters collaboration at least on some level.

Flexibility in Datacop’s Solution

We offer our service as a monthly subscription that you can cancel anytime. With a single subscription, you gain access to all six specialists described in this article, with the smallest package starting at 80 hours per month (equivalent to a part-time commitment).

What truly sets us apart is the ability to reallocate specialist time from month to month. One month, you may need extensive support from an Owned Channels Specialist to craft automated email campaigns. The next, you might shift your focus to data quality assurance, requiring the expertise of a Data Engineer.

If you were to build an in-house team, you would likely need to hire a dedicated specialist for each of these roles. With our solution, you don’t have to. Instead, you can adjust the allocation of hours across specialists as your needs evolve—offering a level of flexibility that’s incredibly difficult to achieve with a traditional team structure.

We act as an extension of your internal technical marketing team, ensuring you get the right expertise, exactly when you need it—every month of the year.

Book a call

Marketing Techstack

It’s also crucial to highlight that the tools these specialists work with make a significant difference. Our extensive experience with various marketing tech stacks has shown that there are huge differences in tool capabilities and the speed at which certain use cases can be deployed.

We’ve seen real-world examples where the same use case took just 2 hours in one setup but required 40 hours in another—a staggering difference.

That’s why we believe choosing the right technology is critical. The tools you select will have a huge impact on the number of hours required from each specialist when deploying use cases. A well-chosen tech stack can significantly enhance efficiency and reduce operational overhead.

In this article, we shared our insights on different approaches to marketing tech stack infrastructure. We outline the key pros and cons of each setup and explain why we favor one approach in particular. We hope you will find this article useful!

For a deeper dive into the tech stack we typically recommend, check out this article:

Use-Case Repository & Prebuilt Solutions

One of the key advantages of our services is our repository of 100+ technical marketing use cases across the areas of our expertise.

In addition, Datacop offers ready-made solutions for many common challenges faced by eCommerce companies. These prebuilt solutions significantly reduce delivery time while ensuring that the first version is built on industry best practices.

One example is our Revenue Loss Prevention Notification System, which you can learn more about in this article:

If you found this post valuable…

We hope you found this article valuable. If so, please consider subscribing (for free!) to receive updates on our latest publications.

Subscribe now

Additionally, if you'd like to discuss any aspect of technical marketing in more detail, feel free to book a meeting with us using the link below:

Book a meeting

About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

At Datacop we have a lot of experience setting up advanced automated email scenarios such as “Low-Stock”, “Back-In-Stock”, “Price-Drop”. These scenarios consistently deliver incremental revenue gains. Learn more about how we implement Back In Stock scenarios at Datacop in a previous substack article here:

However, a common obstacle to the implementation and performance of these scenarios are issues with product inventory mismatch with demand. For example consider when a popular Product A’s inventory is “Back-In-Stock”, such as 25. Because the product is popular, there are 3000 potential recipients of the Back In Stock email with Product A. Assuming a conservative 2.5% conversion rate from the 3000 potential recipients - we could expect to sell 75 stock of Product A from the send out, way more than the 25 available.

This is not desirable, because this may result in a lot of interested customers clicking through to the Product A they were alerted is Back In Stock, however by the time they visit the shop the product has already sold out again.

To deal with this problem a common request we receive at Datacop is - how is it possible to “throttle” sending of the 3000 emails so that we notify only 40 email recipients per stock available at once?

• So that instead of sending an email about Product A being back in stock to all 3000 recipients on day one, only send to a (25 units * 40 rate) = 1000 recipients in day one.

• On day 2 check again how many units are left from Product A. Let’s say that after day one, only 15 units of Product A remained, then (15 units * 40 rate) = 600 recipients on day two. And so on.

How to implement a limit rate per available inventory in Bloomreach?

To implement the limiting rate per inventory solution in Bloomreach Scenario’s we recommend using Customer Limit node. That is one of the available operators in the scenario’s drag and drop menu. A customer limit node sets the maximum number of customers that can pass through that node. Once the limit of the node is met, every additional profile will not be allowed to pass through.

To make this node limit users based on which product_id they are eligible to receive and also to check against the stock_level of the product_id in the catalog, three additional steps are required.

Step 1 - Enable drill down in customer limit node

By default, the customer limit does not distinguish between profiles and counts every profile that passes through the node towards the set limit threshold.

If we want to be able to limit the sends per product_id , you must enable “allow drill down group”. This functionality will add the “drill down group (optional)” box. Here jinja can be placed to add multiple limits within a single node. This will mean that not all customers are counted towards the total threshold count, but each customer contributes only to the limit bucket to which they belong.

The jinja objects that can be used in the drill-down group box are - event trigger, customer attributes and webhook responses.

Step 2 - Define dynamic product_id in drill down

“Back-In-Stock”, “Low-Stock” and “Price-Drop” may have multiple methods of how they can be implemented, typically the use-case generates a trigger event for each customer per time interval (usually a day), that contains within the product_id information used to then populate the email creative.

The trigger events, is what we can use to create drill-down groups within the limit node using the jinja - {{ event.product_id }}.

Step 3 - Define dynamic limit

So now that we can drill down per product_id within a single limit node, how do we retrieve the current stock level count of the product_id? And then where do we place the constant value that defines the # of recipients per stock_level per product_id?

To do the above logic, instead of placing a number such into the “limit to # of customers” box in the node - place the following jinja:

{{catalogs["catalogname"].item_by_id(limit_group).inventory|int*40}}

• where [“catalogname”] refers to the name of the catalog that contains the stock_level information

• where “(limit_group)” is a Bloomreach Engagement special jinja object that refers to the value per customer in the “drill-down group” jinja. So in this case the event.product_id.

• where “.inventory” refers to the inventory attribute name in the catalog

• where “40” is the constant value to define the rate of recipients per stock available

How many emails per unit stock should you send?

When discussing this use-case with clients, sometimes there is a tendency for the number of emails send per unit stock set to anywhere between 2 to 5 to emails per inventory per product per day.

We believe this is way too low.

The reason for this is that the conversion rate of email campaigns to purchase simply is not that high. When there is a rate of 2 emails sent per product per stock, that assumes that 50% of the Back in Stock notification emails will convert to a purchase.

In our experience the automated “Back-In-Stock”, “Low-Stock” and “Price-Drop” scenarios tend to have a conversion rate to purchase from email delivered between - 2.5% to 10%, depending on their settings and the type of product sold. Those are still very high conversion rates to purchase as far as email campaigns are concerned.

The best way to approach this is to monitor the conversion rate to purchase of your Back In Stock campaign and adjust the rate of emails per stock depending on the performance of the campaign.

If you have not implemented it before and thus don’t have a starting CR% base to go from, I would suggest to start with 20 per unit (suggesting that 5% of recipients will convert to purchase) and evalaute after the first 30 days based on the performance whether adjusting the rates is needed.

If the CR% is higher, for example 10% then lower the number of units to 10; alternatively if the CR% is lower, for example 4% then increase the number of units to 25 per stock per day.

Implementing Automated Scenarios

If you’d like to discuss the implementation of automated scenarios in Bloomreach with us or any questions related, feel free to book a free 30 minute meeting below.

Book a Meeting

If you found this article useful

If you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now

About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

Cart abandonment email campaigns were among the earliest examples of personalized communication in eCommerce. When a customer showed intent to purchase an item by adding it to their cart but didn’t complete the checkout, businesses—if they had consent to contact the customer—would send an email. These emails serve as both simple reminders about the abandoned item and as opportunities to offer additional incentives, such as discount codes, to encourage purchase completion.

This type of campaign remains highly effective. Even today, cart abandonment emails are among the best-performing automated campaigns for many eCommerce businesses.

For customers who provide their phone number, SMS messages can complement the email sequence. For those without email or phone contact details, retargeting via paid channels is another option, although it’s becoming increasingly expensive.

While cart abandonment emails are common, today we’ll discuss a less-explored approach: cart abandonment website experiences. This concept, while promising, hasn’t gained much traction yet. One reason might be that implementing email campaigns is often easier than deploying website personalization strategies.

However, advancements in eCommerce technology over the past decade have made solutions like cart abandonment website experiences more accessible than ever.

In this article, we’ll demonstrate a simple yet effective cart abandonment website personalization use case. We implemented it on a dummy eCommerce store using only Bloomreach Engagement—without touching the website’s source code.

Let’s dive in.

Cart Abandonment Hero Image

One of the most effective ways to explain a concept is to show it in action. So, let me walk you through how this use case works. In the GIF below, you’ll see that I’m initially greeted by a generic hero image on the website. However, this image changes dynamically once I add an item to the cart. Take a look:

As shown in the GIF above, the initial hero image is a standard one displayed to most website visitors.

However, after adding an item to the cart and returning to the homepage, the hero image updates dynamically. It now highlights the item in the cart, includes the product name, and features an updated CTA button that directs the customer straight to the cart, making it easier to complete the purchase.

Additionally, we added a small message to create urgency by showing how many units of the product were sold in the past 30 days.

Target Audience

Based on our analysis, approximately 50% of customers who abandon products in their cart return to the website within 30 days.

In fact, depending on the size of your email subscriber list, this group may exceed the audience eligible to receive cart abandonment emails. This is because not all cart abandoners have given consent to receive emails. This reveals a compelling opportunity: utilizing website personalization for this segment can be a powerful driver of incremental revenue.

Final Thoughts

This use case pushes the boundaries of what’s possible in eCommerce, redefining the shopping experience for a new era—one where personalization is truly meaningful and tailored to customer needs.

We believe that the full potential of personalization in eCommerce is yet to be realized, and we’re eager to be part of shaping that future.

If you’re interested in implementing this or similar personalization strategies, we’re confident that we’re the right partner for these initiatives. Not only do we rely on proven tactics that consistently deliver results, but we also embrace innovative ideas that have yet to be fully tested in real-world scenarios—like this one.

We have the expertise to implement these strategies as efficiently as possible, evaluate their impact, and calculate their added value. With our resources, we can handle the process from start to finish.

If your company is committed to innovation and turning tomorrow’s standards into today’s reality, we’d love to work with you. Click the button below to get in touch with us.

Book a Meeting

About the author:
MarTech consultant & Digital product architect with experience leading marketing projects, trusted by clients such as GTO Wizard, Eyerim, Forbes and others.

While human error is a common source of costly mistakes, unexpected external events can also catch businesses off guard. These include changes in user behavior, sudden spikes in traffic, or unforeseen technical issues. These events, whether positive (e.g., a viral campaign) or negative (e.g., a sudden drop in engagement), require immediate attention to prevent revenue loss or to capitalize on emerging opportunities.

That’s why our Revenue Loss Prevention Notifications don’t just monitor for human mistakes; they also alert you to anomalies in your metrics caused by unexpected events. Whether it’s a problem to fix or a chance to scale, your team will always be prepared to respond.

Real Stories: When Mistakes Slip Through the Cracks

1. A Consent Change Nobody Noticed

One of our clients made a small change to their checkout consent text. Overnight, email signups dropped by 70–90%. The new wording, while legally compliant, wasn’t customer-friendly. It took months before anyone realized what had gone wrong, and by then, thousands of potential subscribers and the associated revenue were lost.

While we can’t share exact figures, we can demonstrate the potential impact of this change using illustrative numbers. If the average subscriber generates €25 over 12 months and you’re losing 1,000 subscribers daily, the revenue loss quickly adds up.

• Daily: 1,000 lost subscribers = €25,000 in annualized value.

• Monthly (30 days): €750,000 in lost potential revenue.

• Annually: €9,125,000 in cumulative impact.

Once again, these figures are illustrative and based on example numbers, not actual data. The actual impact will depend on your subscriber value and the exact loss rate.

These numbers highlight just how significant subscriber churn can be.

How notifications would help: If our system had been in place, the steep drop in signups would have triggered an alert instantly. The team could have spotted the issue and rolled back the change within hours, not months.

If you’re interested in a detailed exploration of this issue, check out the dedicated article: How Can an Innocent Text Modification in the Checkout Impact Subscription Rates?

2. The Hidden Cost of Simplifying Checkout

Switching from a three-step to a one-step checkout seemed like a smart move. But the change meant customers didn’t opt into newsletters early in the process anymore. As a result, the client experienced a sharp drop in new email subscribers, which went unnoticed for months.

This issue is explored in detail in the article The Hidden Risk of One-Step Checkout.

Let’s use another illustrative example. If the annual value of a subscriber is €13.45 and the business loses 1,400 subscribers daily, the impact becomes significant:

• Daily Loss: 1,400 subscribers = €18,830 in annualized value.

• Monthly (30 days): Nearly €565,000 in missed revenue.

• Annually: A staggering €6.9 million in cumulative impact.

This demonstrates how even a seemingly minor change can snowball into substantial revenue loss if not identified and addressed early.

Note: These figures are illustrative and based on example numbers, not actual data. Actual impact depends on subscriber value and the exact loss rate.

How notifications would help: A sudden drop in email opt-ins would have triggered a warning. The team could have acted quickly to adjust the checkout process, avoiding months of lost subscriber growth.

3. Cart Tracking Gone Wrong

For approximately three months, a client’s cart_update event failed to trigger for half of their customers. This glitch resulted in many cart abandonment emails not being sent, leading to a significant drop in recovered revenue.

While we can’t share exact figures, let’s use an illustrative example:

• Daily Impact: If a brand experienced 7,000 lost cart abandonments per day, and each cart abandonment email was worth €1.50, this equates to €10,500 in lost revenue daily.

• Three-Month Impact: Over three months, this would amount to an estimated €945,000 in lost revenue.

Note: These figures are illustrative and based on example numbers, not actual data.

How notifications would help: With our anomaly detection monitoring in place, the system would have immediately flagged a significant drop in cart_update events. As soon as these events were firing far less frequently than usual, an automated alert would have been sent to the team. This early warning could have enabled the team to identify and resolve the issue right away, preventing months of lost revenue.

In this case, the absence of a notification system meant the problem went undetected for too long, leading to significant financial losses.

Two Types of Notifications to Stay Ahead

Our notification system is designed to solve two critical challenges:

1. Promptly detecting system failures when key processes stop working.

2. Identifying anomalies—both positive and negative—by analyzing daily KPI performance against recent historical trends.
   

1. Real-Time Alerts: When Things Stop Working

Our alert system triggers the moment something critical breaks. It continuously monitors the inflow of key events, and if no events are detected within an expected time frame (e.g., 30 minutes), the system immediately identifies and notifies the responsible person.

Flexible Thresholds

Our system allows you to:

• Set custom thresholds for each event.

• Adjust thresholds for different times of day (e.g., during nighttime, longer intervals for event detection may be acceptable).

What We Classify as Critical Breaks

Critical issues typically involve tracking disruptions or failures in automated campaigns. For example:

• Cart abandonment emails not being sent.

• The checkout process failing to complete.

• Key events like cart_update not firing.

Real-World Example

Imagine no cart abandonment emails are sent for 2 hours. Our system would immediately send an alert via Slack, email, or SMS to your team, enabling you to take swift action and minimize potential losses.

2. Positive and Negative Anomalies

These alerts identify unusual daily anomalies—both positive and negative—in monitored metrics. Thresholds for each metric can be customized individually and adjusted over time (e.g., detecting a 50% deviation above or below the recent historical average).

Example:

• If revenue or traffic drops significantly below the average for a typical Tuesday, you’ll receive an alert to investigate.

• If there’s a sudden spike in conversions, you’ll be notified so you can act quickly and capitalize on the opportunity.

These thresholds and other settigs are stored in a dedicated Bloomreach profile, where they can be easily adjusted. The profile also acts as a logging system for any defined suspicious events that may occur.

Recommended Metrics To Watch For Positive & Negative Anomalies

Drawing on our experience, we’ve identified key metrics worth monitoring. These metrics enable early detection of potential issues and help uncover emerging opportunities.

How Anomalies Are Flagged

Anomalies are identified by comparing yesterday’s metrics to historical averages for the same weekday over a defined period (e.g., the past eight Mondays). This approach accounts for normal fluctuations, such as seasonality or weekday-specific patterns, ensuring a baseline that is both accurate and relevant.

• Positive anomalies are flagged when a metric significantly exceeds its average—say, by more than 50%—highlighting opportunities like a surge in traffic or conversions.

• Negative anomalies occur when values fall below expectations, helping you quickly identify and address potential issues.

This method equips you with timely, actionable insights, allowing you to resolve problems promptly or capitalize on emerging opportunities with confidence.

Flexible & Transparent

Our system is designed to seamlessly integrate into your team’s daily workflows while providing full visibility and control.

• Seamless Integration: Notifications are sent directly to the tools your team uses every day, including Slack, Microsoft Teams, email, or SMS for urgent situations.

• Centralized Logging System: All notifications are archived in a dedicated Bloomreach profile, offering a transparent record of anomalies. This allows you to monitor trends, revisit past incidents, and identify opportunities for improvement.

Detailed Insights: Each alert includes direct access to comprehensive reports, providing clarity on what happened and why, enabling swift, informed decisions. For instance, in the example below, sessions are broken down by day and traffic source. This allows you to quickly determine whether a traffic anomaly—positive or negative—stems from a specific traffic source or is spread evenly across all sources.

Techstack Flexibility

While the system is designed to integrate seamlessly with Bloomreach Engagement, it’s not limited to this platform. Similar configurations can be implemented using tools like BigQuery and Cloud Functions—a setup we’ve successfully deployed—ensuring adaptability to any technology stack.

Human Errors and Unexpected Events Happen. Identify & Fix Them Before They Cost You Big.

Mistakes and unexpected events are an inevitable part of running a business, but their impact doesn’t have to be. Whether it’s an opportunity, like a sudden viral trend, or a challenge, such as technical issues, these moments can significantly influence your business outcomes.

With Revenue Loss Prevention Notifications, you can detect errors and anomalies in real time, empowering you to resolve problems before they escalate and capitalize on opportunities before they slip away.

Don’t let overlooked details or unexpected events derail your success. Book a consultation with Datacop today and take the first step toward smarter, safer, and more proactive operations.

Book a meeting

About the Author:
Co-founder of Datacop, agency that fulfils marketing operation roles in large eCommerce companies such as OluKai, Melin, Roark, Visual Comfort and Company, Dedoles and others.

The ways customers browse online are constantly evolving. When the internet was first “born”, everyone accessed websites via a Desktop. Today, web-enabled screens come in practically every shape and size and are proliferating into cars, onto our wrists, and (into our glasses?). As the methods for accessing websites evolve, it is essential for marketing and data teams to track this information with greater detail.

The problem is that most session_start events typically only track high-level device categories, such as “Phone / Tablet / Desktop.”

At Datacop, we have found a solution to this problem.

In this article, we will cover the additional attributes that can be added to your session_start event through “Detailed Device Tracking”. We will discuss the results of our internal device testing to assess its accuracy and explore how this data can be applied in practice by e-commerce companies, particularly electronics brands.

“User-agent” is difficult to interpret

Since the early days of the Internet, browsers would add the “user-agent” to their HTTP requests. The user-agent is a string of data passed in the request that enables the server to identify which browser or device is requesting the website, and thus, the server knows which version of the site to serve to the browser or device.

However, there was never a standardisation process in the industry regarding what should be passed on to the user-agent by browser/device combinations. As a result, device manufacturers (such as Apple and Samsung) and browser developers (including Google and Mozilla) would identify their devices and browsers in their own unique ways. Furthermore, competitive pressures among browser developers has lead to at times browsers ‘impersonating’ other browsers in their user-agent to avoid providing their users a poorer browsing experience. Over time, the user-agent has become increasingly difficult to interpret, making it challenging to accurately determine the type of device, brand, or model.

Therefore, session_start / page_visit event tracking by CDPs like Bloomreach Engagement, Segment, and mParticle suffers from 2 limitations when tracking browser and device attributes of visits. Out of the box, they are only able to track high-level and sometimes inaccurate information about the devices and browsers of the sessions.

• Limitation 1: Device attribute values in Bloomreach are limited to a handful high-level values such as: “iPhone”, “Android”, “Other (Desktop/Laptop)”, “iPad”, or “Blackberry”

• Limitation 2: Less prominent browsers, such as Edge (Microsoft) and Yandex, are not recognized by out-of-the-box tracking and are instead misidentified as one of the major browser players, like Google Chrome or Safari.

To overcome these limitations, we have tested a solution for this at Datacop. We recommend implementing a detailed device detection repository in your Bloomreach Engagement / mParticle / Segment SDK.

Detailed Device Detection

We recommend to track the following 5x additional event paramaters to your key front-end events such as session_start / page_visit / view_item / cart_update / etc. :

1. advertised_browser: returns a more accurate browser value

2. complete_device_name: returns the brand name and model of the device, such as “Apple iPhone 14 Pro Max / Samsung Galaxy S23 Ultra”

3. form_factor: returns the device category, such as “Desktop / Tablet / Smartphone / Smart-TV / Robot / Car Screen / Other non-Mobile”. Compared to the standard out-of-the-box device recognition, this recognizes new device types.

4. is_app_view: returns true/false, depending on whether the page was visited via a Mobile App Browser, such as if you open a page via the Instagram / Facebook App.

5. is_robot: returns true/false, depending on whether the session was generated by a bot such as an online crawler or other programmable agent

Testing & Results

At Datacop, we conducted thorough testing of the detailed device detection repository to evaluate its performance in practice.

According to Statcounter, in October 2024, Apple devices represented approximately. 28% of mobile sessions. As you can see on the screenshot below, detailed fingerprinting is capable of identifying the specific make of the iPhone from the 46 potential Apple iPhone models, or iPad from the 38 potential Apple iPad models.

However, we have noted in our testing that if the user has an ad blocker installed on their iPhone, the detailed device detection will not identify the iPhone model, only that it is an iPhone.

According to Statcounter, in October 2024, Android brands such as Samsung and Xiaomi represented approximately. 22% and 10.5% of mobile sessions, respectively. As you can see on the screenshot below, with detailed fingerprinting, it is possible to identify the specific popular Android devices by brands like Samsung S23 Ultra, Xiaomi Redmi Note 11 Pro, and LG G6.

However, we have noted in our testing that not all Android devices are recognisable even with the detailed device detection. Some Android devices are simply identified as “Generic Android 6.0”. Since there are more than 24,000 different Android devices, it is reasonable to have such “catch-all” values for the long-tail of the potential specific Android device brands and models.

Our testing revealed that the detailed device fingerprinting is able to more precisely identify browsers. As mentioned in the introduction, browsers sometimes have incentives to “pretend” to be a different browser. As a result, compared to the standard browser recognition, there are browsers such as Edge, Yandex, and Brave that the detailed device recognition is able to recognize, but are misidentified in the standard device recognition. Note such instances in the screenshot below.

If you click open a webpage on a mobile device within the interface of a mobile application such as Instagram or Facebook, you may have noticed that it can open the session within an “App Browser”. With detailed device recognition, it is possible to distinguish such an App Browser visit. Standard device tracking does not distinguish these sessions from others.

Can you imagine the possibilities for your personalisation and analytics capabilities if you could track the above-mentioned information reliably into your Bloomreach Engagement?

Use-cases

So our testing has revealed that the detailed device tracking can enable your business to track more accurate and detailed device and browser details, especially regarding mobile devices. So, how can this information be used in practice?

1. Personalizing based on Device Model

Detailed device detection enables businesses to craft personalized shopping experiences and marketing campaigns based on the user’s specific device. These use cases are particularly relevant to electronics e-commerce brands that sell or service mobile devices and e-shops that offer accessories for devices.

Device-Personalized Shopping Experiences: Knowing the exact device model allows e-commerce companies to recommend products uniquely suited to the user's needs. For instance, users browsing with an "Apple iPhone 14" can be shown recommended accessories specifically for that model. On category pages, the e-shop may pre-filter on the compatibility with the user’s device model or prompt them to do so. Alternatively, products that are compatible with the user’s device, such as smart watches, laptops, and earphones, could be surfaced to the user.

“compatible with {{ user’s model }}”.

This kind of personalisation can reduce friction for users and improve conversion rates, including adding items to the cart and purchasing on your website.

The personalisation of the user experience using the detailed device can go beyond the web and can be used in retention campaigns. Consider the following examples:

Automated cross-sell campaign for user’s device: send a follow-up email/sms to the user after the session that highlights best-selling services & accessories for that device, such as device covers, device glasses, insurance, etc. Automated up-sell campaign for users’ devices: send a follow-up email/sms to the user after the session that highlights accompanying compatible products with the user’s device, such as laptops, smartwatches, earphones, etc.

Device Upgrade Campaigns: This is specifically relevant for e-shops that sell mobile devices. Identify users with older devices and target them with upgrade incentives. For example, targeted banners on the web could highlight iPhone trade-in programs for users browsing on an "iPhone 11," showcasing the trade-in value and promoting the latest models. Additionally, automatically generated comparison tables between the user’s current device and that brand’s latest model can be used to see what features and capabilities the user is currently missing.

2. Upgraded Analytics on Device / Browser

Enhanced detection provides a more granular view of how websites perform across different devices and browsers. Essentially, in any analysis where the device or browser attribute of the session is concerned, it is upgraded:

Conversion Rate Insights: Analyze whether conversion rates vary significantly by detailed device or browser. Is there a device model that is experiencing significantly below or above average conversion rates in its funnel to purchase? If users on Microsoft Edge have a lower purchase conversion rate than users on Google Chrome, it may indicate a compatibility issue or a suboptimal user experience. The above analysis can be applied to Android vs Apple models, or based on different dimensions of the devices.

Consider comparing the CR% to purchase, Average Order Value, Average Unit Price, and Customer Lifetime Value of customers with different device models. Do users with more expensive mobile devices have a higher average order value (AOV) and a higher customer lifetime value (CLV)? If so, it would make sense to experiment with providing more premium opportunities for users with more expensive devices and perhaps more value deals for users with less expensive devices.

Anomaly Interpretation: Whenever there is a positive or negative anomaly (such as a sudden drop in traffic or a surge in sales), it is always worthwhile to examine the data split by devices to gain a better understanding of what happened. Having more granular device data, as well as more accurate browser readings, would greatly help with interpreting changes in the e-shop.

Customer Understanding: For e-shops that sell mobile devices, electronic devices, or products related to these devices, data on the specific models and brands used by your customers can be a treasure trove of insights useful in many analyses.

3. Cleaner Data & Metrics

Detailed device detection also helps improve the quality of data used in analytics and decision-making by accurately filtering out irrelevant or misleading sessions. Since it is capable of identifying robot sessions more accurately, it provides an additional layer of defense in maintaining data quality.

Unidentified robot sessions and crawlers can skew key metrics, such as traffic, conversion rates, and funnels. By identifying and excluding robot-generated sessions, managers at e-commerce businesses can ensure cleaner and more actionable data.

Implementing Detailed Device Recognition in Your Bloomreach Project

If you’d like to find out how to implement detailed device recognition in your project and discuss how to drive business results with this feature, book a free 30-minute meeting with us at the button below:

Book a Meeting

If you found this article useful

f you enjoyed it, don’t miss out on future articles—subscribe for free and have our latest content delivered straight to your inbox.

Subscribe now