We have another company basically printing money with Google Search.

Another is driving tons of results this quarter from email campaigns to previous users and customers.

And we have another one doing amazing stuff with brand, and they’re seeing a huge lift in organic results because of it.

The point is, there are a million different ways and channels to scale your company. There is no such thing as a universal good channel or a universal bad channel. Only a channel that works or doesn’t work for you.

One of the questions I get asked a lot, usually within the first call, is, “Which channel should we be on?” And to me, that’s kind of a silly question because I have no idea. We’ve probably been talking for 30 minutes. I haven’t gotten under the hood of the business yet. I don’t know enough to tell you that - spoiler alert: nobody does, especially someone sharing a playbook online.

• You need to understand whether your buyers are actually spending time there.

• You need to understand whether the metrics work for your business.

• Can you actually run a profitable ad campaign on that channel?

• Is there enough demand?

• How much does it cost to reach someone?

• How well do those people convert all the way through the funnel?

• How much is a customer worth to you?

All of these things factor into whether a channel is actually going to work.

That’s why it’s so silly when I see people post things like, “Here’s the playbook to scaling your growth.”

There is no playbook.

The only playbook that’s available is the one you build.

You can obviously learn from what other companies are doing. We do that all the time. If I see somebody doing something interesting on Meta, I’m going to look at it. If I see a company getting great results from outbound, I want to understand what they’re doing. But that doesn’t mean I can take exactly what they did, copy it over to another company, and expect the same result.

It doesn’t make any sense.

A company with a $100,000 ACV has completely different economics than a company selling something for $5,000 a year. A company with hundreds of thousands of previous users has an email channel available to them that a startup doesn’t. A company selling something people are already searching for has a completely different opportunity with Google than a company creating a new category.

Even something as simple as the audience matters. Your buyers might spend all day on LinkedIn while another company’s buyers barely use it. So saying “LinkedIn ads work” or “LinkedIn ads don’t work” doesn’t really tell you anything.

They work for who? Selling what? At what price? With what offer?

And guess what? It also changes all the time.

That’s the other part of this that I think gets missed. Even when you figure out your playbook, that doesn’t mean you’ve figured it out forever.

Re engagement campaigns only last so long. If you have a huge database of previous customers and users, you might be able to drive a ton of pipeline from it right now. But you’re eventually going to work through that audience.

Google might be printing money for you today. Then more competitors come in, CPCs keep going up, and eventually the math no longer makes sense.

You might find an amazing Meta campaign that works for months, but eventually the audience gets tired of seeing it or the creative stops working.

SEO changes too. The way people approached SEO five years ago isn’t necessarily the way you should approach it today. has written about this in the context of moving away from thinking about individual keywords and toward building authority around broader topics.

That’s just how these things work. Channels get more competitive. Audiences get exhausted. Platforms change. Costs go up. New channels emerge. Your own company changes.

So I don’t think the job of a growth team is to find the playbook and then just run it forever. The job is to figure out what works for your business right now, push on it while it’s working, and keep testing so you have some idea of where you’re going next.

That’s really the only growth playbook I believe in.

Figure out where your buyers are. Understand the economics. Test things. When something works, do more of it. And when it stops working, don’t keep throwing money at it because somebody told you that channel is supposed to work.

The playbook is different for every company.

And even your own playbook is going to change. The only certainty in growth is uncertainty.

Thanks for reading,

Adam

Every day on Linkedin you see posts about building content engines, launching hundreds of outbound emails, spinning up landing pages in minutes, or generating weeks worth of social content before lunch.

We can debate the quality but the one things that is true is that AI has dramatically reduced the time it takes to build something in marketing.

So we have been lulled into this illusion that because execution is faster, growth should be too. In my experience - that usually isn’t how it works..

Sure, AI has compressed the production cycle, but it hasn’t changed the learning cycle. Those are still two completely different systems operating on different timelines.

• You can create twenty new ads in an afternoon, but you still have to wait for thousands of people to see them before you know which message actually resonates.

• You can launch a new landing page in an hour, but you still need enough traffic to determine whether the positioning is better than the last version.

• You can rewrite your outbound messaging with a single prompt, but you still have to wait for replies, book meetings, and watch opportunities move through the pipeline before you know if you’ve actually improved anything.

Execute all you want but buyers still operates at the same speed they always have.

Customers still need to discover you. They still need to trust you. Sales cycles still take weeks or months. Expansion and retention still take even longer to understand. AI hasn’t changed any of that. I would even argue because there is so much noise out there they actually make it harder.

I’s also created a new problem I am seeing far too regular: mistaking activity for progress.

Teams launch five new campaigns instead of one and assume they’re moving faster. Leadership expects results in weeks instead of months because “AI should make this easier.” Marketing teams feel pressure to constantly ship something new because the cost of creation has become so low.

But the bottleneck in getting growth usually isn’t in the ability to create campaigns, it’s in the ability of Marketers to learn from them and apply the learnings.

And sadly that’s the muscle I think marketing teams are losing. Because AI has made execution so easy, we’re launching more campaigns than ever before. More ads. More landing pages. More outbound sequences. More content. But instead of treating each campaign as an experiment that teaches us something, we’re just throwing more spaghetti against the wall to see what sticks. There is no system to it.

We’re spending less time asking why something worked. Why one message outperformed another. Why one audience converted better than the rest. Why prospects dropped off halfway through the funnel.

Understanding those things and applying them to the next campaign is what actually helps companies grow and scale faster.

If AI saves you ten hours a week on execution, please don’t immediately fill those ten hours with ten more campaigns. Spend them reviewing sales calls. Looking at conversion data. Talking to customers. Documenting what you’re learning and feeding those insights back into the next campaign.

That’s how growth and all of these experiments actually compounds over time. Not because they launch more but because every campaign makes the next one better.

Thanks for reading,

Adam

They’re right.

Dave Gerhardt calls it “the single most important skill to master in business.”

Sam Parr has said, “The single skill that has impacted my business career more than anything was copywriting.”

This isn’t just me saying it.

The best targeting, the best landing page, the best email sequence, and the best ad all fall apart if the copy doesn’t connect.

And that matters more today than ever.

Marketing has never been more competitive. AI lets anyone send thousands of “personalized” emails every day. Buyers are overwhelmed, and generic messaging gets ignored.

So how do you actually stand out?

My friend Eddie Shleyner just launched a copywriting course that I’d genuinely recommend “Transformational Sales Emails”

If you know me, you know I’m usually skeptical of courses. Too many are built by people who teach marketing more than they’ve actually done it. Cash grabs for the seller with little to no value for the buyer.

Eddie is different.

I worked with him for more than two years at G2 and saw his work firsthand. He’s the best copywriter I’ve ever worked with, and more importantly, I’ve seen the impact his writing has on real campaigns.

He doesn’t teach templates or hacks. He teaches the psychology behind why great copy works.

One last thing. This isn’t a paid promotion. Eddie doesn’t even know I’m writing this.

I’m sharing it because I genuinely think it’ll help the people who read this newsletter. When I sat down to write this week’s issue, I couldn’t think of anything more valuable than pointing you toward someone who’s helped shape how I think about copywriting and marketing.

This is the fastest path to becoming a better marketer, I highly recommend checking it out:

https://verygoodcopy.mykajabi.com/offers/ctMsh9Xd/checkout

Thanks for reading

Adam

AI has dramatically lowered the barrier to building software. You can launch products faster, build features in days instead of months, and build an entire go to market motion with a handful of tools.

The problem? Everyone else can too.

We’re entering an era where building the product is no longer the hard part.

Getting noticed is.

At the same time, nearly every acquisition channel has become more expensive or less effective. The old playbooks simply don’t work anymore. The new ones are being written in real time.

• Google Ads costs continue to climb. According to Search Engine Land’s benchmark report, average search CPCs increased nearly 13% year over year, with 87% of industries seeing higher costs. Even if your campaigns haven’t changed, you’re often paying more simply because everyone is competing for the same buyers.

• LinkedIn has become one of the most expensive B2B advertising channels. Recent benchmarks show average CPCs around $11, with many prospecting campaigns landing between $12 and $15 per click. Every wasted impression, broad audience, or mediocre ad has become significantly more expensive.

• Outbound email has become an arms race. AI has made it possible for every SDR to send hundreds or even thousands of personalized emails every week. The average office worker already receives more than 120 emails per day, making attention one of the scarcest resources in B2B.

• Organic search isn’t what it used to be either. Google’s AI Overviews are increasingly answering questions directly in search results instead of sending visitors to websites. Some studies have found AI Overviews appearing on nearly 65% of informational searches, while publishers are already reporting meaningful declines in organic traffic as more searches end without a click.

Every company is fighting for the same attention.

Every buyer has become better at filtering out noise.

And marketers are constantly being asked “Can you grow fast and efficiently?”

• CAC.

• Retention.

• Expansion.

• Margins.

Growth at all costs has become profitable growth at all costs.

So what are you supposed to do?

I don’t think the answer is chasing the next growth hack or hoping another acquisition channel appears. The channels still work. They’re simply far less forgiving than they used to be.

The companies I consistently see winning are investing in four things.

Events

Not giant trade shows with six figure sponsorships.

Smaller, highly targeted events where real conversations happen.

• Founder dinners.

• Customer workshops.

• Executive roundtables.

• Community meetups.

As AI makes digital interactions easier to create, real human interactions become more valuable. Relationships compound in ways algorithms can’t.

Education

This is where I think most companies are getting it wrong. Everyone wants to publish thought leadership.

The problem is that most thought leadership is just another opinion. Buyers don’t need more opinions. They need and want evidence.

• Show me the campaign that doubled conversions.

• Show me the dashboard.

• Show me the experiment that failed.

• Show me what changed after you redesigned the pricing page.

• Teach me something I can actually use on Monday morning.

The companies building loyal audiences today aren’t trying to sound smart. They’re trying to make their customers smarter and are doing it in an honest and authentic way.

Paid with focus

Paid advertising still works incredibly well. What’s changed is that efficiency matters more than scale.

Five years ago you could get away with broad targeting and average creative because clicks were relatively inexpensive. Today, every wasted impression carries a real cost.

The best campaigns I’ve seen over the past year all share the same characteristics. They target a narrowly defined audience, solve one specific problem, and present one compelling offer. Instead of trying to reach everyone, they focus on becoming memorable to the buyers who matter most.

Brand

Ironically, AI may be making brand more valuable than it’s been in years.

When every company can generate decent blog posts, decent ads, decent emails, and decent landing pages, “good enough” becomes the baseline.

Brand becomes the differentiator. People buy from companies they’ve heard of. Companies they trust. Companies that consistently teach them something useful. Companies they remember before they need them.

I don’t think marketing is getting harder because the channels stopped working.

I think it’s getting harder because average marketing has become free.

• AI can create average.

• Templates can create average.

• Your competitors can create average.

The companies that win over the next decade won’t necessarily be the ones spending the most money or publishing the most content.

They’ll be the ones people remember before they ever start searching. The ones that do something better than average.

Happy Friday.

The billboard barely mentioned the actual product. No burger. No fries. No toy. Just:

The takeaway from the post was that B2B companies should be more like Mcdonalds and sell outcomes, not features.

Honestly, that is horrible advice for 99% of start-ups.

What B2B companies can learn from McDonald’s is probably nothing. Your brand is not Apple. It is not Nike. It is not McDonald’s.

Those companies operate with something most B2B companies do not: massive built-in context.

McDonald’s can put up a billboard selling “a quiet car ride” because everyone already understands the mechanics underneath it. Happy Meal. Kids. Toy. Fries. Temporary peace for exhausted parents.

The audience fills in the blanks automatically because the brand literally spent decades and billions of dollars building those associations. If you had never heard of McDonald’s before, that billboard would actually be terrible. You would have absolutely no idea what they sell. Is it a red sound machine? Noise cancelling headphones?

But every time a consumer brand runs an abstract campaign like this, B2B marketers convince themselves the lesson is: “sell outcomes, not features.”

That is not the lesson.

The lesson is that brands with enormous awareness can afford abstraction. Most B2B companies cannot.

This is one of the biggest problems in SaaS marketing right now. Companies try to market entirely through outcomes:

• Increase efficiency

• Unlock growth

• Drive transformation

• Accelerate revenue

• Power the future of work

And the buyer is left wondering: what the f*%@ does this thing actually do?

So everything turns into vague aspiration. Especially in AI.

The Missing Layer Is Mechanism

Strong B2B messaging usually works in this sequence:

Problem → Mechanism → Outcome

Not:

Outcome → vibes → demo request

This is why most AI websites all sound identical right now.

“Autonomous workflows.”
“AI powered operations.”
“Intelligent orchestration.”
“Self-driving execution.”

Cool. What does it actually do? I work in the space and can’t tell you what most companies even do looking at their website.

Because trust in B2B is built through understanding. Especially when:

• implementation risk is high

• budgets are large

• multiple stakeholders are involved

• workflows change

• buyers are skeptical

Which is basically every SaaS category. The abstraction only works after years of repetition and market dominance.

Apple can say “Shot on iPhone.”

Nike can say “Just Do It.”

McDonald’s can sell silence in a car.

Most B2B companies still need to explain the machine before they can sell the dream. And honestly, that is fine.

Clear is usually more effective than clever anyway. My good friend said it better than me:

The core thesis was simple:

AI tools are accelerating vendor discovery, but they are not creating deep vendor conviction.

Buyers are increasingly discovering vendors inside ChatGPT, Perplexity, and Claude first, then going to Google afterward to validate whether those companies actually deserve consideration.

That changes the intent behind branded search dramatically.

Historically, if someone searched a competitor’s brand, the assumption was they were already mostly committed.

Today, that search behavior appears much more exploratory. So we decided to validate whether the data actually supports that theory.

Over the last two weeks, we analyzed competitive search campaign performance across 16 B2B SaaS companies representing just over $1M in combined monthly Google Ads spend.

A few caveats before diving in:

• Spend volume varies significantly across accounts

• Some companies operate in extremely competitive markets while others are in newer categories

• Sales cycles ranged from sub-30-day SMB motions to enterprise cycles exceeding 9 months

• Attribution models differed across accounts

• I’m not a data scientist, and this is not some formal peer-reviewed industry study. This is directional analysis based on real-world account performance patterns we kept seeing repeatedly across the accounts we manage.

With that said, the directional trends were surprisingly consistent across the dataset.

The Biggest Shift: Pipeline Performance

The clearest signal was pipeline generation.

Across the dataset, competitor campaigns generated 1.8x more pipeline per dollar spent compared to the same period 12 months ago.

Not impressions.
Not clicks.
Actual pipeline generated.

Even more interesting, several accounts saw competitive campaigns outperform branded non-competitor search in terms of pipeline efficiency for the first time.

Conversion Rates Are Increasing

The second major shift was conversion efficiency.

Across the accounts analyzed:

• Click-to-demo conversion rates increased 27% YoY

• Opportunity creation rates increased 41% YoY

• Cost per opportunity decreased 18% YoY despite CPC inflation

The surprising part is that CPCs did not materially improve. Average competitor CPCs were actually up 11% YoY across the sample.

Meaning performance improvements were not being driven by cheaper traffic.

The quality of the intent itself appears to be changing.

Buyer Behavior Looks Different

What stood out most across the dataset was not necessarily top-of-funnel engagement metrics.

It was the downstream buying behavior.

Across multiple accounts, buyers arriving through competitor campaigns were:

• More likely to visit pricing pages

• More likely to engage with comparison pages

• More likely to convert into qualified opportunities

• More likely to move into active sales conversations

That’s important because it suggests competitor traffic is behaving less like passive research traffic and more like active evaluation traffic.

CTRs Improved, But Not Dramatically

One thing that surprised us was that CTR improvements were relatively modest compared to downstream conversion metrics.

Across the sample:

• Average CTR increased 14% YoY

• Impression share remained relatively flat

• CPCs increased 11%

In other words, the gains were not primarily happening at the ad-click level. They were happening after the click.

The buyer arriving through competitor search today appears more willing to genuinely evaluate alternatives than they were previously.

The Messaging That Worked Best

The best-performing campaigns had very little in common with traditional “feature comparison” ads.

The strongest performers consistently focused on:

• Tradeoffs

• Operational pain

• Migration complexity

• Category misconceptions

• Scalability constraints

• Workflow philosophy

• “Why teams switch” narratives

Meanwhile, generic “we’re better” comparison messaging underperformed almost universally.

The companies seeing the strongest results were not just intercepting searches. They were reframing how buyers thought about the category itself.

Why We Think This Is Happening

Our working theory is still the same:

AI tools are compressing vendor discovery dramatically.

A buyer asks ChatGPT for “best payroll software” or “best AI security platform” and immediately gets a shortlist.

But the AI response itself is usually not enough to create deep trust or conviction.

So the buyer moves into validation mode.

They Google the vendors.
They compare reviews.
They search pricing.
They read customer experiences.
They evaluate alternatives.

Which means competitor-branded search increasingly represents:
“I’m researching.”

Not:
“I’m decided.”

That distinction changes the economics of competitive campaigns entirely.

I think many GTM teams still underestimate how much AI is reshaping buyer behavior upstream.

Thanks for reading,

Adam

P.S. If you’re thinking through how AI is changing your paid strategy, positioning, or competitive motion, that’s exactly the kind of work we do at Growth Union.

We’re the team behind growth programs for companies like RevenueHero, Plane, Writer, and others, helping B2B teams rethink how modern GTM actually works in the AI era.

If you want a second set of eyes on your paid strategy, competitive campaigns, or overall growth motion, feel free to book some time with us.

If someone searched your competitor’s name, most marketers assumed the decision was already mostly made. The logic was simple: they already know the brand, so they’re probably committed.

That assumption made sense in the old search world. I’m not sure it does anymore.

Because AI is fundamentally changing how buyers build shortlists. And I think it’s quietly creating one of the biggest opportunities competitive marketers have had in years.

Not because buyers are less educated. Because they’re more educated earlier in the process.

A few years ago, the typical B2B buying journey looked something like this:

1. Realize there’s a problem

2. Search Google

3. Read blog posts

4. Discover vendors

5. Build shortlist

6. Compare solutions

7. Book demos

Google sat at the center of discovery. Your SEO strategy, category pages, educational content, and review sites heavily shaped which vendors even entered consideration.

But now a buyer can open OpenAI, Perplexity, or Claude and ask:

“Best payroll platforms for global teams.”

“Top AI security tools for cloud infrastructure.”

“Best lead routing software for inbound demo requests.”

Within seconds, they get vendor recommendations, category summaries, feature comparisons, pricing expectations, implementation considerations, and alternative options.

The shortlist gets built before Google is even opened.

AI is accelerating vendor discovery dramatically. But I think most marketers are misunderstanding what happens next.

Just because ChatGPT recommends a company does not mean the buyer suddenly trusts that company. In fact, I’d argue the opposite often happens.

The buyer receives a list of vendors and immediately goes into validation mode. They open Google. They search:

• “[company] reviews”

• “[company] pricing”

• “[company] vs competitor”

• “best alternatives to [company]”

• Reddit threads

• customer feedback

• implementation complaints

• migration experiences

The search behavior becomes:

“Okay, ChatGPT mentioned this company. Are they actually good?”

That’s a very different mindset than traditional branded search. Historically, branded search often implied: “I already know who I want.”

Now it increasingly means: “I just discovered this company and I’m evaluating whether they deserve to make the shortlist.”

That distinction matters enormously because it means buyers are arriving at branded search highly informed, category educated, actively comparing, but still relatively uncommitted.

And that is exactly why competitive campaigns are becoming more powerful.

Buyer entering Google are often much more open-minded than they would have been in the past. They are not searching because they’ve decided. They are searching because they are trying to decide.

That creates a much larger interception opportunity for competitive campaigns.

Over the last six months, we’ve consistently seen competitive campaigns perform better across multiple accounts once spend reaches a meaningful threshold. Not necessarily because CPCs dropped. And not necessarily because competitors weakened.

The bigger shift is that the nature of the search itself changed.

The person searching the competitor brand today is often earlier in conviction, deeper in research, more comparison-oriented, more skeptical, and more willing to explore alternatives.

The branded search is no longer a declaration of loyalty. It’s part of the evaluation workflow.

We’re currently analyzing a larger cross-account dataset on competitive campaign performance because we started noticing this pattern repeatedly over the last six months.

Early indications suggest competitor-branded search traffic is behaving differently in the AI era:

• More comparison-oriented

• Less brand-committed

• More willing to evaluate alternatives after initial AI-driven discovery

We’re planning to publish the full findings and data breakdown next week once we finish analyzing the data.

I think many GTM teams still operate under an outdated assumption that discovery happens in search and comparison happens later. But increasingly, AI handles discovery instantly. Which means the actual battle moves downstream into validation, differentiation, positioning, narrative, proof, and comparison framing.

More findings to come next week.

Thanks for reading,

Adam

It is that they are deeply obsessed with something outside of work.

Photography.

Woodworking.

Running.

Cooking.

Surfing.

Music.

Something that forces them to think differently.

That might sound unrelated to business performance. But there is actually strong research behind why this matters.

Psychologists call it psychological distance. When your brain steps outside the domain you normally operate in, it becomes better at forming connections across unrelated ideas and experiences. That process activates what researchers call divergent thinking, which is one of the core mechanisms behind creativity and problem solving.

A well-known study published in the Journal of Experimental Social Psychology found that people thinking about problems from a more distant or abstract perspective produced significantly more creative solutions. The brain literally becomes better at generating novel associations when it is not trapped inside the same immediate context.

There is also the concept of incubation.

When you stop actively working on a problem and shift your attention elsewhere, your brain continues processing the problem in the background. That is why ideas appear during a run, while cooking dinner, or halfway through a long walk.

Research suggests this happens through something called the default mode network, a system in the brain that becomes active when you are not focused on a specific task. That network is heavily associated with idea generation, reflection, and connecting distant concepts together.

Which, when you think about it, is exactly what good marketing requires.

Marketing is not fundamentally about tools, despite what LinkedIn might make you believe right now.

It is about perspective.

The problem is that many people in marketing spend all of their time consuming the exact same inputs as everyone else.

The same newsletters.

The same LinkedIn posts.

The same podcasts.

The same frameworks.

The same AI prompts.

Over time, their thinking starts collapsing into the same patterns as everyone around them. That is why so much marketing today feels interchangeable.

But people who spend meaningful time outside that ecosystem often develop stronger and more original thinking.

A photographer learns how framing changes perception.

A woodworker develops patience and precision.

A runner understands endurance and pacing.

A musician develops rhythm and emotional timing.

A chef learns balance, sequencing, and experimentation.

None of those things are marketing directly. But all of them shape how someone sees problems. They create new mental models. They help people recognize patterns others miss. They introduce analogies and perspectives from completely different worlds.

That is where many of the best ideas actually come from.

Not from staring harder at the same problem everyone else is staring at.

But from stepping far enough away from it that your brain can return with a different lens.

So… go get a hobby other than vibe coding.

Thanks for reading,

Adam

If you spend enough time on LinkedIn right now, it feels like every marketer is suddenly becoming an engineer. People are building internal AI tools with Claude Code, spinning up automations over the weekend, connecting workflows together, and proudly posting screenshots of systems that replaced hours of manual work.

Some of it is genuinely impressive.

But the reality I see talking to 20 to 30 companies a month looks very different.

Most of what is being built inside marketing teams today is happening in silos.

One person builds an AI workflow for content creation. Another creates a lead scoring automation. Someone in RevOps builds routing logic nobody else fully understands. A growth marketer connects five tools together with Zapier and hopes it keeps working six months from now.

Individually, none of these are bad ideas. The problem is that most of them are disconnected from the larger operating system of the business.

Over time, companies end up with:

• Fragile automations that require constant maintenance

• Inconsistent data flowing between systems

• Reporting nobody fully trusts

• AI workflows built around outdated messaging

• Operational logic that breaks the moment the company changes direction

• Teams dependent on one person who originally built the workflow

Eventually leadership will start asking a frustrating question:

How did we automate so much and somehow create more complexity?

The answer is simple.

Building one-off workflows is not the same thing as building scalable systems.

Right now there is a growing assumption that every marketer should become deeply technical. That the future marketer is part strategist, part operator, part engineer.

I do not think that is entirely true.

Should marketers understand AI tools? Absolutely.

Should they become more operational? Yes.

Should they understand how automation impacts workflows and customer experience? Without question.

But acting like every marketer should be architecting scalable systems is unrealistic. Those are completely different skill sets.

A strong marketer typically excels at things like:

• Positioning / Messaging

• Customer psychology

• Creative strategy

• Audience understanding

• Demand creation

• Storytelling

• Market insight

• Channel expertise

A systems-minded operator thinks differently. They focus on:

• Workflow architecture

• Operational dependencies

• Data structures

• Integrations

• Lifecycle logic

• Automation frameworks

• Scalability

• Infrastructure reliability

Sometimes those skill sets overlap. Often they do not.

And when companies blur those lines too aggressively, things start breaking.

For years, marketing teams were constrained by production capacity. The challenge was execution.

Could you create enough content?
Could you launch enough campaigns?
Could you execute quickly enough?

AI changes that equation dramatically. Now almost everyone can produce more. Which means the advantage shifts somewhere else.

The companies that win over the next few years will not necessarily be the ones producing the most content or automating the most workflows. They will be the companies that build the cleanest systems around strong strategic inputs.

That means investing in things like:

• Clear messaging architecture

• Clean lifecycle stages

• Reliable attribution frameworks

• Thoughtful lead routing systems

• Unified reporting structures

• Operational visibility across the funnel

• AI workflows connected to actual business logic

This shift has changed how we increasingly work with clients. A growing amount of our work is no longer just campaign execution or content production.

It is helping companies design operational marketing infrastructure. The layer underneath modern growth teams that connects:

• CRM architecture

• AI workflows

• Lifecycle automation

• Lead routing

• Data enrichment

• Reporting systems

• Outbound orchestration

• Content operations

Because without that operational layer, most AI adoption inside marketing teams becomes nothing more than a good linkedin post.

Lots of activity. Very little compounding leverage.

This is where I think the idea of “marketing engineering” becomes genuinely important. Not as another trendy LinkedIn job title.

And not as “everyone should learn to code.” But as a real operational function inside modern growth teams.

Someone has to think through how the entire system works together.

Questions like:

• How does data move across the funnel?

• How do AI outputs connect into downstream workflows?

• How does reporting remain trustworthy as systems scale?

• How do automations evolve as the company changes?

• How do we reduce operational drag instead of adding more tooling chaos?

Most companies do not need more disconnected automations. They need systems that support marketers instead of overwhelming them. Systems that allow them to focus on the inputs which is where they can have the most leverage and impact.

Because marketers should still spend the majority of their time doing things humans remain uniquely good at:

• Understanding customers

• Refining positioning

• Building trust

• Creating differentiated messaging

• Developing strong creative angles

• Identifying market shifts

• Asking better strategic questions

The goal should not be turning every marketer into an engineer. The goal should be building systems that allow marketers to operate at a higher level.

That is a very different thing. And honestly, I think this becomes one of the defining differences between growth teams over the next few years.

Some companies will use AI to generate more noise. Others will use it to build better systems.

The second group will compound much faster.

Over the last few months, this has quietly become a much bigger part of the work we are doing at Growth Union.

Not just demand generation or campaign execution. We’re increasingly helping companies architect scalable marketing systems across AI workflows, CRM infrastructure, lifecycle automation, reporting, routing logic, and operational design.

Basically: marketing engineering for modern growth teams.

If your team is struggling with disconnected automations, messy lifecycle logic, unreliable reporting, or AI workflows that do not scale cleanly, we should talk.

Book a call with us and we can walk through where your current systems are breaking down and what a scalable marketing infrastructure could actually look like for your team.

Thanks for reading,

Adam

Hanging in the front hall of my house is a print by photographer Harold Feinstein called Blanket Toss.

It was taken in 1955 on the beaches of Coney Island.

A group of kids hold a blanket stretched tight while another kid flies wildly through the air above them. Everyone in the frame is doing something different. Some are focused on catching him. One kid is laughing directly into the camera. The whole thing feels chaotic and joyful at the same time, I love it.

Feinstein grew up around Coney Island and started taking photographs there as a teenager with a borrowed camera. He used to describe it as a “photographer’s paradise” because there was humanity and energy everywhere you looked.

What’s interesting is that this photo was taken nearly 70 years ago. Today, your iPhone could technically take a “better” photo.

• Higher resolution.

• Better low-light performance.

• AI stabilization.

• Perfect exposure adjustments happening in real time.

• Auto focus.

• Burst mode.

The technology is infinitely better. But that does not make you Harold Feinstein. And I am willing to bet there is nothing in your camera roll even close to as good of an image like this.

Why? Because what makes this image memorable has nothing to do with the camera. It is the story. It is the feeling.

It is the ability to recognize something deeply human unfolding in front of you and frame it in a way that makes other people feel something too.

That feels a lot like what is happening in marketing right now. AI has democratized execution. Anyone can generate content. Anyone can spin up campaigns. Anyone can build landing pages, automate workflows, create ads, or pump out hundreds of articles.

Which means the differentiator is shifting is not the tech. Just like in photography, there are a lot of people becoming obsessed with the gear.

• The newest AI tools.

• The newest workflows.

• The newest automations.

• The newest prompts.

But tools alone do not make great marketing. The same way a better camera does not automatically create a meaningful photograph.

The companies that will stand out over the next few years will not necessarily be the ones with the biggest AI stack. They will be the ones that know how to tell better stories.

The ones that understand tension. Emotion. Human behavior.

The ones that know how to communicate an idea in a way that people actually remember.

Because technology always evolves.

The camera changes.

The software changes.

The tools change.

But the ability to make someone care? That part stays timeless.

And I think a lot of companies are about to learn this the hard way. By the end of this year, many teams will realize they produced more marketing than ever before and still missed their goals.

More content.

More campaigns.

More outputs.

But none of it stood out.

Because when everyone has access to the same tools, the winners are usually the people who have something worth saying in the first place.

Thanks for reading,

Adam

Underneath all of those questions is a bigger one they’re really trying to answer: what does a marketing team even look like in an AI-first world?

So here is my honest take:

Most marketing organizations are still built around execution. The assumption is that growth comes from producing more. More campaigns, more content, more output.

That logic made sense when execution was the constraint.

It’s not anymore.

I’ve talked about this a ton already in the newsletter but execution is becoming abundant. You can spin up campaigns faster than ever, generate endless variations of content, and move from idea to launch in a fraction of the time it used to take. And yet, most teams aren’t getting better. They’re just producing more.

Because the constraint didn’t disappear. It moved from execution to decision-making and system design.

Once you understand that, the structure of the team starts to change.

The two roles everything collapses into

If you strip it down, marketing teams are collapsing into two core functions: strategists and builders.

Strategists decide what should exist. Builders make it real.

That sounds simple, but in practice, these are completely different skill sets.

A strong strategist is not just someone with ideas. They have judgment, they have taste. They can take a messy market, unclear positioning, and scattered inputs and turn it into a clear direction. Turn it into a story and a brand that stands out. When they hand something off, it is obvious what needs to be built and why it should work.

You can usually tell the quality of a strategist by how much interpretation is required after they speak. If the team leaves with more questions than clarity, the strategy isn’t done.

A strong builder thinks in systems, not tasks. They don’t just launch campaigns, they design how campaigns get launched. They connect tools, define workflows, and make sure that once something works, it can run again without starting from scratch.

The easiest way to spot a weak builder is that everything depends on them. The easiest way to spot a strong one is that the system runs without them constantly intervening.

As much as everyone wants the opposite to be true - these aren’t the same person. Sure there might be some crazy unicorn out there, but the best storytellers, brand builders are rarely the best operationally and vice versa.

Where performance actually comes from

Most teams blur these roles together and that’s where things start to break.

You get strategists who are pulled into execution, so they never think deeply enough about direction. Or you get builders who are asked to make decisions they don’t have the context for, so they default to whatever is easiest to ship.

The result is a lot of activity without real progress.

What actually drives performance is the interaction between the two.

The strategist defines a clear direction. Not a list of ideas, but a point of view on what should win and why. The builder translates that into a system that can execute consistently. Once that system is live, both roles stay in the loop. The strategist refines direction based on what’s working. The builder tightens the system so it runs faster and with less friction.

Over time, the loop gets tighter. Ideas move faster, feedback comes quicker, and the system compounds.

The structure looks more like a feedback loop than a pipeline:

Direction → system → output → feedback → iteration

Most teams are still operating like a pipeline. That’s why they feel slow even when they’re busy.

How you know if this is working

This is also where measurement starts to change.

Strategists shouldn’t be judged by how much they produce. They should be judged by how often their thinking leads to something that actually works. Are the ideas translating into performance? Is the direction getting clearer over time? Is the team moving faster because decisions are easier to make?

Builders shouldn’t be judged by how busy they are. They should be judged by how much leverage they create. Are campaigns getting to market faster? Is there less manual work over time? Does each new campaign get easier to launch than the last? Is the system improving automatically from what we are learning.

If both roles are doing their job, you start to see a few things happen:

• Time from idea to execution shrinks

• Fewer things get launched, but more of them work

• The system improves with each cycle instead of resetting

That’s usually the signal that you’re building something that compounds.

Where this breaks

Most teams don’t fail because they misunderstand the model. They fail because they over-index on one side.

If you lean too far into strategy, you end up with good thinking that never ships. The team spends more time discussing ideas than executing them.

If you lean too far into building, you get efficient execution of the wrong things. Systems get more complex, but results don’t improve.

And then there’s the assumption that AI replaces both roles. It doesn’t. AI makes good strategy scale faster and good systems run more efficiently. It doesn’t decide what matters or how things should be built.

There’s also a timing issue that shows up as companies grow. Early on, one person can play both roles because speed matters more than structure. At some point, that breaks. If you don’t separate the roles, everything bottlenecks around a few people. If you separate them too early, you slow down before you have enough signal to guide decisions.

Most teams get that transition wrong.

What this means going forward

This isn’t just a hiring change. It’s a shift in how you think about your growth engine.

Your advantage is no longer your ability to produce more. It’s your ability to decide what’s worth building and to build systems that improve over time.

Most teams are still asking how to do more marketing.

The better question is how to build a system that does marketing for you.

When execution is effectively free, the only thing that compounds is good decisions and the systems that carry them forward.

I could also be completely wrong here and we all end up building a machine that ends up taking over the world while we try to find meaning in our new roles as humans.

Thanks for reading,

Adam

Pipeline slows down, and the instinct is immediate. Spend more on ads. Hire more reps. Add more volume at the top of the funnel. The assumption is simple: if we just push harder, we’ll grow faster.

That might work if you can outspend your competition. Most teams can’t.

So they end up stuck in the middle. Not spending enough to dominate attention, but not differentiated enough to stand out. Just adding to the noise.

There’s another way to win, but it’s not new. It’s just more relevant than ever.

The idea of “out-teaching” your market has been around for a long time. The best companies have always done it - look at Drift, Gong and countless others. They didn’t just promote their product, they shaped how people understood the problem in the first place.

But I think it matters even more in 2026. Content, ads, landing pages, outbound, all of it has never been easier to produce. AI has collapsed the cost of creation. Anyone can generate something that looks polished and sounds credible.

Which means most of what’s being produced is converging toward the same average. More content is being created, but less of it is actually useful.

And that’s what makes this moment different.

The advantage is no longer in production. It’s in perspective.

The System: What Actually Compounds

When you try to outspend, you’re buying attention.

You’re renting distribution, layering spend on top of spend, hoping volume makes up for a lack of clarity. It can work, especially in the short term. But it has a ceiling. It scales with budget, and the moment you pull back, performance follows.

Out-teaching works differently.

Instead of trying to capture attention everywhere, you shape how people think before they ever enter a buying process. You define the problem in a way that resonates. You give language to something they’ve felt but couldn’t articulate. You connect dots they hadn’t connected.

By the time they start evaluating solutions, they’re already operating inside your framework.

At that point, you’re not just another option. You’re the reference point.

What We’re Actually Seeing

Across the companies we’ve worked with, the pattern is consistent.

Teams that rely heavily on spend tend to have fragile systems. Higher acquisition costs. Lower conversion rates both in ads and in sales cycles. Constant pressure to maintain volume. Messaging that shifts frequently because nothing is anchoring it.

Teams that invest in teaching behave differently.

Their content gets referenced, not just consumed. They see more direct traffic and stronger first-touch conversion. Sales cycles shorten because buyers show up with a clearer understanding of the problem.

It’s not that they aren’t spending. It’s that their spend is supported by clarity.

Where This Breaks

This approach isn’t universal.

If you’re in a true commodity market with little differentiation, teaching alone won’t carry you. If you don’t actually have a unique perspective, you won’t teach anything meaningful. You’ll just repackage what already exists.

And this is where most teams fail.

They think they’re teaching, but they’re summarizing.

The other constraint is time.

Outspend can generate pipeline quickly. Out-teach builds leverage more slowly. It requires consistency and conviction, which is why most teams abandon it before it starts to compound.

Or they never get leadership buy-in, so 2 months into this approach they have their CEO asking “Where are the leads?!”

Practical Implications

If you actually buy into this, it should change how you operate.

1. Shift from output to insight

More content is no longer the goal. Better thinking is. If your content could be generated without your input, it won’t differentiate you. The bar is whether it changes how someone understands the problem.

2. Build around topics, not assets

Your blog, ads, landing pages, outbound, all of it should reinforce the same core ideas. Most teams create disconnected assets. The ones that win build systems of ideas that show up consistently across every touchpoint.

3. Invest in original perspective

This is the hardest part. It requires forming a point of view based on real experience, not just aggregating best practices. The question isn’t “what should we write about,” it’s “what do we believe that others aren’t saying clearly.”

4. Align your GTM around teaching

Marketing can’t do this alone. Sales, product, and leadership all need to be bought in and reinforce the same narrative. If marketing is trying to teach one thing and sales is pitching another, the system breaks.

5. Use spend to amplify clarity, not replace it

Paid channels still matter. But they work best when they distribute a clear idea, not when they try to brute-force attention. Spend should accelerate understanding, not compensate for a lack of it.

• I’ve seen this first hand repeatedly - paid performs exponentially better when it’s attached to a point of view. When your brand is known for something, your ads don’t have to do all the work. When you’re unknown in a sea of sameness, they do. Hello high CPA.

Outspending buys you attention. Out-teaching earns you trust.

And in a world where everyone can produce content, trust is the only thing that compounds.

Is it easy? No.

Does it work? Yes.

But as the old saying goes, if it were easy, everyone would be doing it.

Thanks for reading,

Adam

Daniel Pink calls these regrets of inaction. The things that stay with people are not the risks they took. It is the risks they talked themselves out of.

There is a direct parallel in growth. Most companies are not struggling because they made one catastrophic bet. They are slowly losing because they never made the obvious hard one.

They never:

• repositioned early

• shifted budget aggressively into what was working

• killed a channel everyone knew was underperforming

• rebuilt the funnel instead of layering more on top

• narrowed their ICP

• leaned into a new distribution surface before it was proven

They waited until the decision felt safe. By then, the upside was already gone. That is the growth version of regret of inaction.

How the System Actually Works

Growth does not compound from constant motion. It compounds from a small number of asymmetric decisions made early enough to matter.

Most teams operate like this:

• optimize what already exists

• add incremental channels

• test creative variations

• build more reporting

This creates activity, but not inflection. The inflection usually comes from decisions that feel disproportionate at the time:

• shifting from keyword-first to topic-first authority building

• collapsing a broad ICP into a much narrower one

• reallocating budget away from “safe” channels into concentrated bets

• changing the sales motion instead of adding more leads

• turning founders or operators into the primary distribution engine

These are not small optimizations. They are system-level changes. And they feel uncomfortable because they force tradeoffs.

You cannot do everything. You cannot serve everyone. You cannot hedge every outcome.

So teams delay. They ask for more data. They extend timelines.They add more stakeholders and more meetings.

What looks like discipline is often just delayed accountability.

The AI Shift Changes the Timeline

This dynamic is about to get more extreme. AI is not just another channel or tool. It is creating entirely new surfaces for advantage.

New ways to:

• build and distribute content

• personalize funnels

• automate parts of GTM systems

• rethink hiring and team structure

• reshape what a “product” even is

We are already seeing early versions of this:

• Teams replacing entire workflows with AI-assisted systems

• Founders becoming primary media channels with AI amplification

• Lean teams outperforming larger ones because they move faster

• Products bundling services, automation, and intelligence together

These are not edge cases. They are early signals. Over the next few years, a series of windows will open. Some will be short-lived. Some will define entire categories.

And the same pattern will repeat: Early movers will look wrong, inefficient, or premature.

Then it will work.

Then it will become obvious.

Then it will become crowded.

Then it will become expensive.

If you wait for proof, you are entering at the most competitive point in the curve.

Evidence and Patterns

We have seen this pattern across multiple shifts:

Content and SEO: The companies that built topic authority early dominated for years. Once everyone moved to “content at scale,” the advantage compressed.

Paid acquisition: Early adopters of Facebook and LinkedIn ads saw outsized returns. As more budget entered, CAC increased and performance normalized.

Product-led growth: What started as a differentiated motion became expected behavior across SaaS.

In each case:

The value accrued to the companies that moved before the playbook was obvious. Not the ones who executed it most perfectly. Internally, we see the same thing with clients.

The biggest gains rarely come from:

• improving conversion rate by 10 percent

• testing another ad variation

• adding one more nurture sequence

They come from decisions like:

• rebuilding routing and speed-to-lead systems

• consolidating messaging into a single narrative

• focusing the entire GTM motion around a tighter ICP

• cutting half the channels to double down on what is working

These are the decisions teams tend to delay the longest or worse they never make them.

Failure Modes and Tradeoffs

This is not an argument for reckless decision-making. There are real risks:

Move too early, and you can:

• invest in the wrong category or channel

• build systems your team cannot support

• create internal chaos without clear direction

There is also a difference between:

• making a bold, informed bet

• making random, disconnected bets

The goal is not to say yes to everything new. The goal is to recognize when something matters before it is universally validated.

Another failure mode is mistaking motion for conviction. Many teams “experiment” with new ideas in a way that removes all real risk:

• small budgets

• no real ownership

• no structural changes

This protects downside, but it also caps upside. You cannot get asymmetric outcomes from half-committed bets.

Practical Implications

If you take this seriously, the question changes.

It is no longer:

“What is the safest next step?”

It becomes:

“Where are we clearly under-indexing because the decision feels uncomfortable?”

A few ways to think about it:

1. Identify the delayed decisions: Where has the team been saying “we should probably do this” for the last 3 to 6 months?

2. Separate risk from visibility: Are you avoiding the decision because it is wrong, or because it is highly visible if it fails?

3. Look for early signals, not consensus: If something is already widely accepted, most of the upside is gone.

4. Commit at the system level: If the bet is real, it should change how resources, time, and attention are allocated.

5. Accept imperfect information: You will not get full certainty. Waiting for it is often the riskier path.

Most leaders assume their career will be defined by the bets that failed.

More often, it is defined by the bets they never made. The same is true in growth. Not every decision should be aggressive. But the companies that win are rarely the ones that waited for perfect clarity. They are the ones that moved while the decision was still uncomfortable.

Because once it feels safe, it is usually too late to matter.

Thanks for reading,

Adam

PS - If you are looking to partner with an agency who can help you find what bets to make and move fast let’s talk. Book a call with me today.

“Human attention is finite and not fungible. When someone gives their attention to something, they are giving a piece of their finite life to it. Important activities deserve full presence, not a fraction of it. The tools we build for those activities should signal: this is not like everything else.”

If you want to read the full piece, it’s worth it:

Todd Curtis

At what point does delegation become self-erasure?

A colleague of mine, someone who actively uses, benefits from, and advocates for AI tools, sighed deeply before telling me…

Read more

7 months ago · 23 likes · 6 comments · Todd Curtis

That idea has been stuck in my head because it cuts straight through what’s happening across marketing right now.

We’ve made production infinite. But attention is still fixed. And we’re acting like those two things scale together, they don’t.

You can see the mismatch everywhere.

Ad creative is spun up in batches. Landing pages get cloned and lightly tweaked for every segment. Blog content is generated at scale to “cover” topics. Outbound messages are written, skimmed, and sent to thousands.

On paper, it looks like leverage. In practice, it shifts the cost somewhere else.

Because none of this reduces the cost of consuming it.

Every ad still asks for attention. Every page still asks for attention. Every piece of content still asks for attention.

And attention is the one thing that didn’t get cheaper.

So the burden moves from the creator to the audience.

We save time by automating. They pay for it by filtering.

Estimates suggest the average person is exposed to between 4,000 and 10,0000 marketing messages (including online, TV, and physical ads) daily. So ask yourself - how many ads do you remember or did you engage with last week? You saw 70,000 of them, there must be hundreds, no? The truth is it is probably a handful if that.

People have gotten incredibly good at filtering. Watch the next time someone is scrolling through the feed and you will see how quickly you have to earn attention.

Not because they sit there analyzing every sentence, but because they can feel it. They can feel when something was rushed, templated, or barely thought through. You don’t need to tell someone something was generated. They know when it’s not worth their time.

This is where most teams get it wrong.

They think the problem is distribution. If we just get this in front of more people, something will hit.

But distribution doesn’t fix something that isn’t worth paying attention to. It just spreads it faster.

You see the same pattern across channels.

In SEO, teams flood the site with content that technically targets the right terms but never builds real authority. It’s the exact trap of surface-level coverage and content that competes with itself instead of compounding.

In paid, creative gets iterated endlessly, but none of it actually earns a pause.

On landing pages, the value is “communicated,” but nothing pulls you in or makes you care enough to stay.

Across all of it, the system is optimized for output, not for earning attention.

The uncomfortable part is that AI didn’t break this. It just exposed it.

A lot of what’s being produced now was already low-effort before. It just took longer. Now it’s faster, cheaper, and a lot more obvious.

If attention is someone giving you a piece of their life, then most of what we’re putting out right now isn’t just ineffective. It’s asking for something we didn’t earn.

The teams that win aren’t going to be the ones producing the most.

They’re going to be the ones who slow down just enough to make it clear that what they made is worth paying attention to.

Because the real constraint was never production.

It was always whether it deserved someone’s time.

Thanks for reading,

Adam

PS: Unlock growth with a hand-picked team of the top marketers in tech.

“A true extension of the team. They just get it. No fluff, no wasted budget, just smart, effective marketing that drives pipeline and revenue.” Mark Huber, VP of Marketing at UserEvidence

Book a call with me to learn more.

Ask most marketing teams about them and you’ll hear the same things. The CPCs are too high. The leads are low quality. The platform works for brand awareness but not pipeline.

Over the past year we analyzed just over $15 million in LinkedIn ad spend across multiple B2B companies. Different industries. Different deal sizes. Different ICPs.

Creative mattered.

Targeting quality mattered.

Offer design mattered.

But one pattern showed up again and again. Most LinkedIn lead generation campaigns start in the wrong place.

Instead of matching the offer to where buyers actually are in the market, companies jump straight to demo requests and product messaging. The logic usually goes something like this: we know exactly who our ICP is, we have the account list, let’s go put demo offers in front of them.

The reason is simple. Companies are treating LinkedIn like search.

But LinkedIn is not search.

When someone types a query into Google, they are actively trying to solve a problem. They already know something is broken and they’re looking for answers.

LinkedIn works differently.

People aren’t logging in to evaluate vendors. They’re scrolling through industry commentary, seeing what people in their network are talking about, keeping an eye on competitors, or hate-scrolling past an endless stream of selfies paired with a hot take about AI.

The mindset is different.

Which means most people seeing your ad are not actively shopping for your product. Best case scenario - they are still orienting themselves around a problem.

When a campaign immediately asks for a demo in that context, it creates an awareness mismatch. The ad is asking for a buying decision before the buyer has even fully framed the problem.

That mismatch explains a lot of the frustration teams have with LinkedIn lead gen.

The targeting might be perfect. The ICP list might be dialed in. But targeting the right companies does not mean those buyers are ready to talk to vendors.

What We’ve Seen Work Instead

Across the campaigns we reviewed, certain types of assets consistently performed better when audiences were cold.

The common thread was simple. They helped buyers understand a problem before asking them to evaluate a product. So what worked?

Industry benchmarks: When companies published data about how their market was actually performing, engagement rates tended to spike. People want to know where they stand relative to their peers.

Teardown-style analysis: Breaking down how a system works, where it breaks, and how companies typically approach solving it tends to draw in operators who are trying to improve their own process.

Tactical resources: Usually after audiences had already engaged with earlier content. Guides, frameworks, and playbooks attracted buyers who already understood the problem and were actively trying to solve pieces of it.

Direct demo offers, on the other hand, tended to perform best only after audiences had already interacted with the brand in some way.

None of this is particularly surprising once you step back and think about how B2B buying actually works. Most purchases start with problem exploration, not vendor selection.

This advice feels pretty basic and you might be reading this thinking the same. But the majority of LinkedIn campaigns just styart with the demo. They start at the point where the buyer is supposed to already trust you when they likely don’t even know your brand or company. So why would they ever want a demo from you?

So Why Does This Keeps Happening?

There are a few reasons companies default to demo-first campaigns. Sales teams want meetings, so marketing optimizes toward the most obvious conversion event.

Targeting tools create a false sense of readiness. When you can upload an account list that perfectly matches your ICP, it feels logical to assume those buyers are ready to talk.

And demos are easy to measure. Earlier-stage engagement is harder to attribute, which makes it harder to justify internally.

Put those forces together and you end up with a system where most LinkedIn campaigns start at the bottom of the funnel.

Which is exactly where most audiences are not.

Thanks for reading,

Adam

P.S. If you’re running LinkedIn ads and they’re not producing the results you expected, and you’d like a second set of eyes on what you’re running today, you can schedule a call with us here.

A company decides they need to “move faster.” They spin up AI content engines, build a few vibe coded calculators, launch a benchmark report, maybe an ROI tool.

Suddenly the team is shipping a lot.

• Ten blog posts a week.

• A new asset every month.

• Endless landing pages.

On the surface it looks productive. But when you step back and ask what actually changed in the business, the answer is usually: not much.

The problem is that most teams are optimizing the wrong side of the equation. They are obsessed with outputs when the real leverage comes from inputs. And AI has made this gap worse.

Generative AI dramatically lowered the cost of producing marketing assets. Articles, landing pages, tools, email campaigns, ad variations. All of it can now be produced at a fraction of the time and cost. Naturally, companies respond by producing more.

More pages.

More campaigns.

More “content engines.”

But if the inputs are weak, the outputs simply scale the wrong thing faster. Weak positioning becomes hundreds of pages of weak positioning. Surface level ideas become hundreds of surface level posts.

You do not get leverage. You get noise.

This is not just anecdotal either -

A Gartner survey found that more than 60 percent of CMOs say their marketing technology stack is underutilized or failing to deliver the expected impact. At the same time, budgets for martech and AI tooling continue to grow.

The issue is not the tools. It is how they are being used.

What Inputs Actually Mean

When I talk about inputs, I am not talking about prompts. I am talking about the thinking that should happen before anything gets produced.

Questions like:

• Who exactly are we selling to?

• What problems actually keep them up at night?

• How do they describe those problems in their own language?

• What objections show up late in the sales cycle?

• Where do they currently go to look for answers?

• How do we stand out or approach things differently?

Most companies skip this part because it feels slower. It requires talking to customers. Looking at lost deals. Analyzing where opportunities stall.

None of that produces something you can screenshot and show in a Slack channel. But it determines whether everything you produce afterwards actually works.

The Differentiation Test

There is another input question that rarely gets asked before launching something.

Does this actually help us stand out?

Look at almost any SaaS category right now.

Every company has a benchmark report.

Every company has a guide to the same topic.

Every company has an ROI calculator.

None of those assets are inherently bad. But if two hundred companies in your category can produce the same thing, it stops being differentiation and starts becoming table stakes.

Before launching any new asset, the better question is:

If two hundred companies tried to create this same thing, would ours look meaningfully different?

If the honest answer is no, the problem is not the output. The problem is the input.

Differentiation almost always comes from the inputs:

• Unique data

• Contrarian viewpoints

• First hand operational experience

• Deep customer insights

Without those inputs, the output becomes interchangeable. And interchangeable marketing rarely wins.

Content strategy is a perfect example of how this plays out.

Most teams still start with a keyword list. Then they create an article for every keyword. Fifty keywords. Fifty posts.

But this approach tends to produce thin, disconnected content that fails to build real authority. The reason is simple - the input was wrong.

Instead of understanding the topic deeply and mapping the real problems buyers care about, the strategy was driven by search volume. Things just don’t work that way anymore, increasingly evaluate authority at the topic level rather than individual keywords.

Which means surface level coverage across hundreds of posts rarely compounds. Depth does.

The Goal Alignment Problem

There is another input issue that quietly breaks a lot of marketing strategies. Targets that do not actually move revenue. You will hear things like:

“We want to publish 40 blog posts this quarter.”

Or

“We want to launch three new tools.”

But those are production targets. They measure activity, not outcomes. The real input question should be: What result actually moves the business forward?

• Pipeline created

• Revenue influenced

• Deals accelerated

• Expansion within existing accounts

Then the next question becomes:

Will this initiative realistically drive that result? A surprising amount of marketing activity fails that test, especially if it was vibe coded over a weekend.

So Why Companies Default to Outputs

Outputs are easier to show.

• You can point to a new landing page.

• You can show a new calculator.

• You can say you published 30 articles this month.

Inputs are harder to show.

Customer research.

Market analysis.

Deep thinking about positioning.

But those are the things that determine whether the next six months of marketing actually work.

The teams getting the most leverage from AI are not the ones producing the most. They are the ones improving their inputs. They spend time talking to customers. They analyze why deals are won and lost. They understand the problems their buyers actually care about.

Then they use AI to scale the outputs.

AI becomes an amplifier. Not a replacement for thinking. And that difference shows up everywhere in the results. Because in marketing, better outputs rarely fix bad inputs.

But better inputs almost always create better outputs.

Toward the end of this year a lot of companies are going to look back and wonder how they missed their numbers despite producing more marketing than ever.

More content. More campaigns. More AI generated assets.

The dashboards will show activity everywhere. But activity is not the same as impact.

If the inputs were wrong, all AI did was help scale the wrong strategy faster.

Thanks for reading,

Adam

I worked with a client who was exceptional at this in the worst way possible.

Before we could launch an email campaign, the immediate question was: what happens if too many people book demos and we cannot support it with just two SDRs?

Before launching paid ads: what happens if people download the asset and we do not have a proper long-term nurture built out yet?

Every initiative turned into a stress test for demand that did not exist. Meanwhile, nothing was live. Weeks passed. Meeting after meeting debating problems that didn’t exist.

This is the pattern I see repeatedly. Teams design for overflow before they have flow.

How does this look in reality:

• Team A ships five new tests every two-week sprint. That is 26 sprints per year, or 130 experiments annually.

• Team B ships five new tests per quarter. That is 20 experiments per year.

Now assume a 10 percent hit rate, which is generous in most B2B environments.

• Team A ends the year with 13 scalable wins.

• Team B ends the year with 2.

That is an 11-win delta in a single year.

But the real gap is not the wins. It is the 117 failed tests Team A ran. Those failures generate signal. They sharpen messaging, eliminate weak audiences, refine positioning, and build intuition about what actually moves pipeline.

Team B debates theoretical constraints. Team A engineers around real ones.

Over three years, the difference becomes structural. One team builds a portfolio of compounding levers. The other builds caution.

So why does this happen?

In my experience, it usually comes down to three things.

1. Fear of making mistakes. Shipping something imperfect feels risky. If a campaign underperforms, it is visible. If a landing page converts poorly, someone has to answer for it. So teams attempt to eliminate risk through planning. But most risk in early growth is informational, not reputational. You are trying to learn. Avoiding small failures delays meaningful insights.

2. Leadership control. In some organizations, every campaign requires executive review. Every subject line gets debated. Every audience needs approval. This creates the illusion of rigor. In practice, it slows iteration speed dramatically. Control feels responsible. Speed looks reckless. In growth environments, that assumption is often inverted.

3. A culture of meetings. Calendars are full. Slides are built. Notes are captured. It feels like work is happening. But meetings are not work. Say it again with me - meetings are not work. Meetings are meant to align before doing the work. The work is the shipped experiment, the launched campaign, the live asset collecting data.

You can spend an entire quarter aligning on edge cases for a campaign that never runs. Meanwhile, another team ships five campaigns, learns from all of them, and reallocates budget based on performance.

There is a difference between operational readiness and fear-based planning.

• Operational readiness asks: if this works, can we respond within reason?

• Fear-based planning asks: what could go wrong at scale, and how do we prevent it before we even launch?

The first creates motion. The second creates delay.

If too many demos book, you hire. If too many downloads come in, you build nurture based on real behavioral data. You do not pre-optimize for demand you have not proven.

Most early growth constraints are imaginary. SDR capacity is not the current bottleneck. Nurture depth is not the current bottleneck. Ad scale limitations are not the current bottleneck.

The bottleneck is shipped experiments.

If your hit rate is 10 percent, your primary lever is volume of quality attempts.

Ship five things every two weeks for a year and you will have 13 scalable wins and a far sharper understanding of your market.

Spend that same year solving for edge cases in meetings, and you will have caution and growth numbers that won’t make you feel great at the end of the year.

Thanks for reading.

Adam

The setup was intentionally long-horizon. The agent was given a starting cash balance. It had to order inventory from suppliers. Deliveries could be delayed. Suppliers could go out of business. Competitors were operating in the same environment. In one variation, multiple AI “operators” ran vending machines side by side, triggering price wars and strategic maneuvering.

The instruction was simple: maximize profit.

In updated versions of the test, one model meaningfully increased its balance across several runs. It adjusted pricing. It negotiated with suppliers. In competitive mode, it coordinated pricing with rivals and effectively formed a cartel to protect margins. It even exploited weaker competitors by selling them inventory at markups.

If you wanted to spin that into something dramatic, you could lead with the headlines:

“AI just killed the CEO role”

“CEO’s are cooked”

But no one did. I didn’t see it once on my Linkedin feed.

Because everyone understands, instinctively, that running a company is not the same thing as maximizing profit inside a constrained simulation.

A CEO is not a vending machine optimizer.

The job includes hiring and firing, shaping culture, allocating capital, setting long-term direction, managing investor expectations, navigating regulation, handling crises, and making tradeoffs between short-term earnings and long-term durability. Profit is an outcome of those decisions. It is not the full scope of the role.

So when an AI agent performs well in a profit-maximization benchmark, we interpret it correctly. It is a signal of technical progress in planning and optimization. Not a referendum on executive leadership.

What’s interesting is where that restraint disappears.

• When a model writes a strong cold email, the narrative becomes “SDRs are dead.”

• When a model generates landing page copy, it becomes “copywriters are cooked.”

• When a model builds ad variations or drafts campaign plans, it becomes “agencies are obsolete.”

The logic is the same in every case. A constrained task improves. A measurable output increases. The role gets reduced to that output. The entire profession is declared cooked.

The difference is that we do not reduce CEOs to a single output metric. We do reduce marketers, sales reps, agencies and others that way.

Marketing has been framed for years as “running ads” or “writing copy.” Sales as “sending emails” or “closing deals.” Agencies as “managing campaigns.” If that were the full scope of the work, the threat would be existential. Those are exactly the types of mechanical tasks that models are compressing quickly.

But those descriptions are incomplete.

Marketing at a meaningful level is about deciding who you serve and how you win. It is about positioning tradeoffs. It is about choosing which segments not to pursue. It is about aligning narrative with product truth and long-term pricing power. It is about deciding when not to optimize for short-term performance because it damages brand or future leverage.

Sales is not just message output. It is reading power dynamics inside accounts. It is mapping influence. It is knowing when a champion is losing ground internally. It is deciding whether discounting a deal strengthens revenue or weakens future negotiations. It’s about building meaningful relationships with people.

Agencies, at their best, are not media buyers. They are external decision frameworks. They pressure-test strategy. They prevent expensive mistakes. They see patterns across companies that internal teams cannot.

These roles are bundles of judgment layered on top of execution.

The vending machine benchmark showed that AI can reason through pricing, inventory, and competition when the objective is clearly defined. That is real progress. But it does not eliminate the need for people who define the objective, design the constraints, and decide which tradeoffs are acceptable.

If we would reject the claim that CEOs are obsolete because a vending machine agent increased profits, we should be equally skeptical when someone declares marketing or sales dead because a model wrote decent copy.

The real shift is not the extinction of professions. It is the compression of their mechanical layers.

Professionals who operate only at the mechanical layer will feel pressure first. Those who operate at the level of judgment, constraint design, and system architecture will not disappear. Their leverage will change.

The vending machine story doesn’t prove leadership is finished. It exposes how casually we confuse a measurable slice of work with the entirety of a role.

And we only seem to make that mistake when the role feels easier to dismiss.

And if I were to put my tin hat on for a moment, I might say this:

No one would dare post “CEOs are cooked” after that experiment if they actually had a CEO.

But it’s very easy to post “marketers are cooked” when you don’t sit in that seat, don’t understand the full scope of the role, and don’t report to one.

Some extinction narratives are about technology. Others are about distance from power.

Thanks for reading,

Adam

P.S. If your current agency feels cooked, we’re not. Growth Union helps teams redesign their growth engines for the AI era without losing the strategy layer. Book a call with us today.

Consideration Is Earned, Not Bought

Wynter recently surveyed B2B SaaS CMOs and asked a deceptively simple question: What gets a vendor into your consideration set?

Respondents had to rank one channel as their primary driver. Here is what CMOs ranked #1:

1. Word of mouth: 42%

2. Brand fame: 18%

3. G2 reviews: 12%

4. AI recommendations: 10%

5. Google research: 8%

6. Supplier content: 6%

7. Paid ads: 2%

8. Cold outreach: 2%

Word of mouth is twenty times more influential than paid advertising in getting onto a shortlist. Meanwhile, paid ads and cold outreach, the two channels that absorb the majority of B2B marketing spend and operational effort, barely register.

This is not a creative problem or a targeting problem. It is a systems problem.

The Mistake Most Teams Make

Most companies assume inconsistent pipeline is a distribution issue. If performance is flat, the instinct is to optimize creative, tighten targeting, increase budget, or launch a new outbound motion.

But the Wynter data suggests something more fundamental. Consideration is formed before a buyer ever clicks your ad.

By the time someone engages with your LinkedIn campaign or downloads your whitepaper, they have already filtered you through a set of unspoken questions:

• Have I heard of this company before?

• Do people I know and trust use them?

• Would I look competent bringing this into my organization?

Word of mouth answers those questions upstream.

Brand familiarity reinforces them.

Reviews validate them.

Paid ads show up later.

Which leads to the line most teams need to internalize:

Paid ads don’t create consideration. They amplify it.

The conclusion is not that paid ads do not work. The conclusion is that paid works best when there is something credible to amplify. If your ads feel inefficient, it may not be a media buying issue. It may be that there is no underlying trust engine doing the heavy lifting.

Consideration Is a Credibility Loop

Consideration does not behave like a linear funnel. It behaves like a credibility loop. In practice, it unfolds something like this:

1. A peer mentions your product in conversation or online.

2. Your brand appears repeatedly in category discussions.

3. Reviews and commentary reinforce a consistent narrative.

4. AI systems and search results reflect that consensus.

5. When your paid ad appears, it feels familiar rather than risky.

Most B2B teams start at step five and attempt to brute-force the rest through budget. That is backwards and usually ends in horrible results.

Familiarity lowers friction. Unknown brands, regardless of CPC efficiency, struggle to convert.

AI Is Formalizing Word of Mouth

One signal in the data deserves special attention. 10% of CMOs ranked AI recommendations as their primary driver of consideration. That now outranks Google research according to this.

Two years ago, that number would have been effectively zero. AI is not replacing word of mouth. It is formalizing it.

Large language models synthesize patterns across reviews, mentions, articles, and public conversations. If enough people describe your company in similar ways, the model surfaces that consensus. If no one talks about you, AI has nothing to aggregate.

This is why “optimizing for AI” without building real social proof is misguided. Visibility in recommendation engines is an output of credibility, not a shortcut to it.

So what are we supposed to do? Just sit there and wait until someone talks about us?

Designing Word of Mouth Intentionally

If word of mouth is the most powerful driver of consideration, the obvious pushback is:

So what are we supposed to do? Just sit there and wait until someone talks about us?

No.

You cannot force organic advocacy. But you can design the conditions that make it far more likely and far more frequent.

Here are three structured ways to approach it.

1. Turn Customer Wins Into Public Signals

Most companies collect testimonials. Few convert them into distribution assets.

Instead of burying praise on a case study page, create repeatable mechanisms that move customer outcomes into public channels. Ask your customers to post about you on social media.

A simple structure:

• Ask one focused question tied to a real outcome: “What changed for you after implementing this?”

• Offer a guaranteed incentive for participation.

• Provide light posting guidance so language reinforces your positioning.

• Secure permission to reuse those posts as paid creative.

If 10% of a 1,000-customer base participates, that creates 100 peer-authored endorsements circulating inside relevant networks. Those posts influence buyers organically, feed AI consensus signals, and give paid media something credible to amplify.

This is not about bribing praise. It is about converting real outcomes into visible signals.

2. Build Advocacy Into Your Operating Rhythm

Word of mouth compounds when it is systematic, not episodic.

Instead of running one advocacy campaign per year, embed prompts into natural business moments:

• Onboarding completion

• Measurable milestone wins

• Expansion or renewal conversations

• Quarterly business reviews

Advocacy should feel like a byproduct of success.

The key metric here is not just total review count. It is velocity and recency. Fresh signals suggest momentum, and momentum drives consideration.

3. Design for Amplification, Not Persuasion

If paid ads amplify consideration rather than create it, your job is to design assets that are worth amplifying.

That means prioritizing:

• Customer-authored LinkedIn posts

• Public endorsements

• Review excerpts

• Screenshots of peer commentary

Then use paid media to distribute those signals into tightly defined ICP audiences.

This is also why LinkedIn thought leadership ads are performing so well right now. When buyers repeatedly see people inside a company talking intelligently about the exact problems they face, it builds familiarity and trust long before a sales conversation happens.

It humanizes the brand and signals expertise in the problem space, not just the product space.

By the time a direct-response ad shows up, the company is no longer a stranger.

Paid media becomes dramatically more efficient when it distributes credibility instead of attempting to manufacture it.

Where This Breaks

This approach is not a shortcut.

It breaks if the product does not generate meaningful outcomes. It struggles in markets where no shared narrative can form. It disappoints leadership teams that expect immediate pipeline impact from upstream investments.

Word of mouth compounds slowly, then suddenly. Early signals are subtle. The payoff is delayed.

But once credibility crosses a threshold, amplification channels such as paid and outbound become far more productive.

The Strategic Question

If 42% of CMOs say word of mouth gets a vendor considered and only 2% say paid ads do, most B2B budgets are inverted.

The real question is not how to optimize paid acquisition in isolation.

It is whether your company is systematically increasing the number of credible people talking about you in relevant contexts.

Solve that, and paid likely becomes dramatically more efficient. Ignore it, and no amount of optimization will compensate.

Consideration isn’t something you can just buy. It is something you have to earn and then amplify.

Thanks for reading,

Adam

Getting answers has never been easier.

If you work in B2B marketing, you no longer need to spend weeks researching how something works at a conceptual level. You can ask how to stand up outbound, structure an ABM program, design a RevOps stack, or plan experiments, and get a coherent response almost instantly.

That is a real step forward. But marketers are actually getting better because of it or worse?

The problem is not that people are using AI to learn. The problem is that people are arriving at conclusions without having to wrestle with the process that produces them. They have the answer but don’t have the understanding of what went into it.

The gap usually becomes painfully visible when something goes wrong.

Where the Gap Actually Shows Up

When systems behave as expected, surface-level knowledge looks like competence.

Campaigns launch > Dashboards update > Leads come in > Nothing forces deeper thinking.

But B2B marketing systems rarely start out or even stay stable for long.

When outbound response rates drop suddenly, people who learned outbound through frameworks tend to focus on copy tweaks or subject line testing. People who have built and rebuilt outbound programs are usually one step ahead of this before it happens because they know the warning signs. They monitor domain health, deliverability, list decay, sequencing logic, timing, and saturation. They know which failure modes appear most often and which signals matter early.

When paid acquisition becomes more expensive, the surface response is to blame the platform. More experienced operators monitor whether the audience is exhausted, whether creative is being over-rotated, whether intent has shifted, or whether the offer still maps to where buyers actually are in their decision process.

When site traffic drops, the default reaction is often to publish more content. Operators with experience step back and ask whether the site has real authority, topical cohesion, original signal, and a clear reason to exist beyond rephrasing what already ranks.

None of these issues show up in theoretical AI plans. They only surface once you are operating inside a live system with real constraints, real budgets, and real consequences.

Why AI Makes This More Subtle

AI is extremely good at explaining ideal states. It summarizes best practices, outlines frameworks, and gives clean answers to messy questions.

What it cannot provide is the kind of understanding that forms when something breaks and you do not immediately know why.

That understanding is built through exposure to failure. It comes from seeing the same problems appear in slightly different forms, learning which variables matter in practice, and developing intuition about what not to touch.

Judgment is not knowing what usually works. It is knowing how to respond when the usual thing stops working or staying ahead of what you know are the typical pitfalls.

AI can help people reach a reasonable starting point faster.

It cannot remove the work required to build judgment.

The Commoditization Problem No One Talks About

There is another risk in relying too heavily on AI-generated plans that every seems to just ignore… If everyone is starting from the same baseline answers, no one has an advantage.

AI tends to produce sensible, average starting points. That is useful for avoiding obvious mistakes, but it also means teams converge on similar structures, similar sequences, and similar playbooks.

Real advantage comes from knowing how to move beyond the baseline. That requires experience.

It shows up in decisions like:

• Which experiments to run first and which to defer

• How to stack tests so learnings compound instead of conflict

• When to simplify before adding complexity

• Which common mistakes are not worth making because they waste time and signal

• How to run a differentiated campaign form a creative perspective.

Those decisions do not come from reading or prompting. They come from having built systems that failed in predictable and unpredictable ways.

Two teams can start with the same strategy. The one with experience will move faster, avoid dead ends, and adapt as conditions change. The other will follow the plan until it breaks and then struggle to diagnose why.

This is why experienced operators I know are not worried about being replaced.

AI can generate plans, but it cannot prioritize tradeoffs under pressure.

It can explain systems, but it cannot feel where they are fragile.

It can suggest experiments, but it cannot sequence them based on lived constraints.

As markets become more competitive and systems more interdependent, judgment becomes more valuable, not less.

AI changes the starting point. It does not change what mastery requires.

We are entering an era where access to answers is universal.

The advantage will belong to the people who still earn judgment.

Thanks for reading,

Adam

PS If you want to shortcut theory without skipping the hard-earned judgment, this is where bringing in an experienced agency actually matters. At Growth Union, we have helped scale teams at G2, Writer, UserEvidence, Help Scout, and others by turning plans into operating systems and helping teams know what to do when things stop working, not just when everything looks good on paper.