This video is about one diagram I badly needed two months ago.

At the time, I was building for clients what people refer to as a GTM “second brain”. The same thing as a harness - if you come from the world of engineering. A point of confusion kept creeping into client convos... Then I realized it’s about the same point of confusion I had, back it the day.

What was going on

I’d read about a cool tool, build something with it, leave it half-connected, then start another thing. Agents, webhooks, scheduled jobs, memory files, scripts... stuff everywhere.

323 commits later, I had built a lot.

I had also created a big friggin mess.

The tools weren’t the problem. I was treating “AI GTM” as one big thing when I was actually building three different systems at the same time.

Once I separated them, the mess started to make sense and I found a better way to think about or frame “building GTM systems”. Before this, when I heard the term, a swirl of repos, agents and openclaws would emerge in my mind.

In most people’s minds, too. So a simple separation helps:

Subscribe now

Workflow complexity - the stuff that gets work done. Skills, plugins, agents…

Autonomy - how a capability (skill, plugin, agent) runs without you.

Memory and evals - how the system gets smarter.

Sounds boring, I know. But an ancient Chinese proverb attests:

(Pretty sure it’s 3000 year-old wisdom.)

The video explains the model, but this is the part I couldn’t fit into eight minutes: what those three systems look like in real work, where I’ve screwed them up, and how I’d build them now.

System 1: What can the workflow actually do?

Most people start as prompt tourists.

You open ChatGPT or Claude, paste in some company context, get a decent result, close the window and do roughly the same briefing again next Tuesday.

Useful? Absolutely.

A system? Nay.

The first real step (01 in the ladder pictured above) is almost embarrassingly boring: a shared context repository. A folder containing your ICP, buyer roles, jobs to be done, alternatives, positioning, product information, voice rules and the stuff your team currently keeps scattered across decks, docs and somebody’s head.

I’ve now built these for several companies. One client’s repository captures the ugly way its buyers solve the problem today: logs in one tool, tickets in another, user context somewhere else, then a human manually reconstructing what happened.

That one description feeds the ICP, buyer personas, customer signals, competitor analysis and positioning. The content workflow reads it. The outbound workflow reads it. The sales material reads it.

Nobody has to rediscover the company before doing the work.

That’s Level 1. Files. Shared truth. No sexy demo required.

Then come skills.

A skill is a repeatable instruction set that reads the context and does one job. Apply my voice to a draft. Research a company against the ICP. Turn a messy brain dump into LinkedIn hooks. Take a client call and pull out the evidence that matters.

This newsletter uses one. It reads my voice rules, newsletter examples, content strategy and source material, then moves through angle, outline, substance, rewrite and polish.

I still trigger it. I still decide whether the result is any good. But I don’t explain from scratch how I write a newsletter every time.

From there, workflows get wider.

My weekly content digest currently watches 84 sources across LinkedIn, newsletters, YouTube, Reddit, X and news sites. It doesn’t just summarize 84 people. That would be a fairly expensive way to create a worse inbox.

It looks for topics appearing across at least three independent sources, separates convergence from noise, maps the useful ones to my content pillars and keeps a rolling 15-week history so “rising” means something beyond “the AI noticed it today.”

Another workflow takes raw client calls through transcription, evidence extraction, strategy synthesis and then into a client-ready deck. Several tools. Several steps. Still a mostly linear journey from messy source to usable output.

At the more complex end, the workflows branch and start writing back into systems other workflows use.

For one client, we built a multichannel outbound coordinator across Smartlead and HeyReach. A Cloudflare Worker holds the real state of each prospect. If somebody replies on either channel, it records the event, cancels every unsent action and pauses the other channel. If a provider fails halfway through an enrolment, the workflow stops for review instead of blindly trying again and possibly sending twice.

At that point, you’re building infrastructure, not asking a chatbot to write an email.

System 2: How much freedom should it get?

Workflow complexity and autonomy get shoved together all the time. They’re not the same.

You can have a five-step workflow that waits for you to press the button. You can also have a cron job that runs by itself every Sunday and produces complete nonsense.

The cron job is technically more autonomous. Congratulations, I suppose.

Autonomy is simply how much freedom the system gets to start, decide and act without you.

So a workflow isn’t better per se if it’s more autonomous… but more autonomy is the general goal - if you’re following me :)

At the lower end, you trigger the work but let AI make choices inside it. My newsletter workflow can suggest angles, choose what to cut and restructure a weak section. It cannot wake up on and decide by itself that I owe the internet a newsletter and publish whatever it fancies.

Taste-heavy work stays close to the human.

The content digest is different. It can run on a schedule because the consequence of a mediocre digest is low. It lands in Slack, I read it, and nothing gets published, nothing outward-facing gets produced.

Then you get event-triggered workflows.

In one client system (in-person training events), a new registration starts a chain that calculates the right course, first class start date, fee and recipient type, then chooses from six email templates and prepares the complete message.

But it stops there.

A human sees the actual recipient and full email, then checks an approval box. Change the source data or template after approval and the approval disappears. The email provider accepting the request (hi Brevo) is tracked separately from the email actually being delivered.

There is plenty of autonomy inside that workflow. There is also a deliberate human gate before the part you can’t just casually undo.

That’s the design question I needed to learn to ask more often. I’d default to, “Can AI run this without a human?” - which isn’t the wrong question per se, but I learned to think deeper about that.

The better question is, “Which decisions can it make safely, and where would one mistake become embarrassing, expensive or illegal?”

For the outbound systems we build, webhooks can route the next action automatically. A reply can stop both channels automatically (Linkedin and Smartlead). A redelivered webhook cannot create a duplicate touch.

But the system refuses to enrol the first person until a human has approved the exact campaign revision, copy, timing and list.

High autonomy where the rules are clear. Friction where judgment or consequence is high.

At the far end you get specialist agents assigning work to other agents, monitoring systems and running loops, with plenty of caveats.

Most companies aren’t there.

Most don’t need to be yet.

You earn more autonomy by proving the workflow works with less autonomy first. If it breaks while you’re watching, you sure as hell ain’t giving it more freedom. Yet.

System 3: Does it get less stupid next time?

This is the part I underestimated most and frankly, it’s where the greatest room of improvement lies once you have a relatively complex repo and workflows going.

Memory begins at the bottom. Your ICP file and other context are actually memory. So are the examples inside a brand-voice skill and the list of claims your sales workflow isn’t allowed to make.

But storing information is easy-peasy and doesn’t get you far.

The system needs a way to decide whether an output was good, understand why it failed and carry the useful correction into the next run.

Evals, in other words. Checks against a standard.

Here’s an example I used to tell:

Early on, I scraped 76 company careers pages looking for marketing job openings (a common signal source). The workflow presented 17 results.

Hand checked a few. And then all 17. Zero were jobs.

Service pages. Testimonials. Team bios. Newsletter CTAs. The scraper saw marketing-related words and treated the whole internet like a careers page.

I wasted time checking a completely dead prospecting pipeline.

Ooops. My mistake.

The honest diagnosis: I had defined the job badly. The workflow didn’t know what counted as a real job posting, so it returned keyword matches and I acted surprised.

I fixed the process at the time. But the “permanent guardrail” I later claimed would prevent the failure forever doesn’t even exist anymore.

Which rather proves the point.

Leave a comment

A lesson sitting in my head isn’t system memory. Neither is a rule in a file nobody loads. If the correction isn’t attached to the workflow, tested and retained, the system can make the same dumb mistake six months later with a newer model.

The content digest that tells me what is trending, has lightweight memory and evals.

• It keeps 15 weeks of topic history.

• During the first few weeks it isn’t allowed to call something a trend because there isn’t enough baseline yet. Links must come from the collected source data rather than being reconstructed from memory.

• Three independent sources count as convergence. One mention is just one mention.

The client email system has a different kind.

• An API response saying “accepted” does not count as delivery.

• The workflow waits for the provider’s webhook before changing the status to delivered.

• If source copy changes, the old human approval no longer counts.

The outbound coordinator goes harder again. It records an event before acting on it. Every action has an idempotency key (sorry for the jargon, just wanted to sound fancy) = so the same webhook can’t create the same action twice. An ambiguous failure goes to review, not an automatic retry.

Same idea, but with different consequences: define what good looks like, capture what happened, compare the two, then change the next run.

At the top end, lessons shouldn’t be able to promote themselves into permanent rules just because an agent had a bad day.

In the memory system I’m building now, a new lesson starts as a candidate. It needs evidence, an evaluation and human approval before it becomes a rule other workflows automatically load.

Otherwise “self-learning” can quickly become “one weird result rewrote the operating system.”

Not ideal.

One brain, three very different builds

Put the three systems next to each other and the maturity question changes.

My newsletter workflow has decent complexity, low autonomy and strong human evaluation. That’s needed, because voice and taste are super important.

The content digest has more scheduled autonomy, plus enough history to tell a recurring signal from this week’s noise. It informs a human. It doesn’t publish.

The outbound coordinator has branching logic and event-triggered actions, but also the hardest approval, deduplication and stop rules because mistakes reach real people.

What to do now that you know this.

More autonomy isn’t automatically more mature.

More complexity isn’t automatically more useful.

And without memory and evals, both eventually create entropy: more moving parts, more mistakes, less trust, until nobody wants to touch the bloody thing.

That’s the mental model I wish I’d started with.

If you’re currently fiddling with AI for marketing or sales, don’t begin by asking which agent framework you need.

Start here:

1. What repeatable job are you trying to improve?

2. What context does it need to do that job consistently?

3. How will you know whether the output is good?

4. Which parts can run safely without you, and where should the workflow stop?

5. What happens to the lesson when it gets something wrong?

Answer those and the tooling gets a lot easier.

Ignore them and you can spend 323 commits building a very advanced machine that nobody, including you, fully trusts. (Ask me how I know.)

I know which route I’d take now.

What are you building - and which of the three systems is currently lagging behind?

Reply and tell me. Also tell me if you prefer less-complex GTM AI stuff, more complex, or if this is just right…

Want to know where people are, so I can keep this Substack relevant for the majority of my beloved readers.

Dan Renyi,
Klear

Today’s newsletter’s job is to help fix your story and product messaging. It also has zero copywriting or messaging advice, because copywriting ISN’T how you begin to fix most copy problems.

Most messaging problems are grounded in the way your ICP (idea customer profile) is defined.

Most companies try to define their ideal target company starting with industry, size, and tech stack. That's the wrong order.

I get why. Open Sales Navigator, ZoomInfo, Apollo, LinkedIn Ads - every single one of them is built the same way. Industry. Headcount. Tech stack. Geography. Then, roles. So when it's time to define who you're actually selling to, your brain reaches for the doc that matches the tools you’re used to. Makes sense. Doesn't work.

That's targeting, friend. Not strategy

Firmographics and all those easy-peasy filters tell you who to reach but ofc don't tell you what to say once you get there. And when you don't know what to say, messaging goes broad - because broad feels safer than accidentally excluding someone who might pay us a dollar.

I used to think messaging is product’s/wordsmithing’s problem (aka copy). To figure out what to say to an audience once we have narrowed them down. Then I had this hunch that a really good ICP should double as a value proposition.

Sounded like a dumb idea at first, and then it began to make sense. If you want to know the behind-the scenes on how I arrived there (otherwise skip to the next section):

1. there is the “the list is the message” thinking in cold outbound circles that says your list should be so well targeted that you basically don’t need personalization in the message, bc the list’s filtering carries the specificity.

(Firmographics still belong in the exercise, to be clear. Just not at the start.)

2. there is the old wisdom from Fullfunnel.io rockstars that ICP should center around use cases or jobs-to-be-done (JTBD). If you think about it, it rhymes with the first idea.

Then Anthony Pierri had a good post that also brought up the “ICP=JTBD” idea and added that you could technically build the ICP definition around a product category if that makes more sense.

So for the past 2-ish years, I was subscribing to the idea of putting jobs to be done in the ICP definition.

Now the thing about most jobs to be done frameworks is that they are qualitative. And usually lack KPIs and OKRs. There’s a good reason for that which we can’t go into right now But it started to make so much sense to really drill down into the actual KPIs that our solution impacts - and connect them with the actual responsibilities of your buyers.

Again, usually only copywriters start to inquire about that as they’re building product pages and collateral, so that they can put numbers like “23x-ed client’s conversion rate” on decks.

What I have found is

we need to figure out the KPIs we impact way before it comes time to put pixels on a web page.

Here's the better starting point

I call it SUcK.

Situation, Use case, KPI.

I wish I had found a more positive acronym, but it works and it doesn’t need to be cute.

Every serious senior B2B buyer has a dashboard. On it: one or two numbers that decide whether they keep their job. Not twenty metrics. One or two. If those stay red long enough, they get fired. Their career is riding on a short list of indicators, and most of them know exactly which ones.

Your job as a vendor is to get as close to one of those numbers as you can, and prove you move it. Nail that, and you're mission-critical. Leave the connection fuzzy - even by one or two logical steps - and you land in the nice-to-have pile.

That isn’t the pile you wanna be in.

Process: use SUcK to build an ICP that doesn’t

Five steps, mapped under those four letters. Let's make this concrete with a hypothetical illustration: a field service software SAAS (AI-powered, above all) - sold to home services companies. HVAC, plumbing, electrical.

Step zero is to get very clear and aligned on your solution’s capabilites, strengths, weaknesses… but you knew that.

1. Situation: find the business context when the need arises for your solution’s category. Maybe the HVAC client's team-size grew past a critical point where pen-and-paper or spreadsheets don’t cut it any more. Or they may have acquired another company and now things are more complex. Find the typical and less-typical buying triggers.

2. Find the mission-critical Use case. Not helpful. Not useful. Mission-critical - the job where, if you weren't there, someone would notice fast. For field service, that's not "organize work orders." It's "dispatch the right technician to the right job so the business gets more revenue out of every working day."

3. Find the champion closest to that use case. Here, it's the VP of Operations. Not the CEO. Not the CFO. The VP Ops lives and dies by dispatch performance specifically.

4. Find the business-level KPI closest to that use case. In home services, there's usually one number on the wall that matters most: revenue per technician per day. It's what leadership asks about Monday morning. It's what the VP Ops gets bonused on - and fired over.

5. Find the process-level KPIs you actually move, that roll up to that number. Almost everyone skips this step. A vendor doesn't move "revenue per tech per day" directly. What they move is jobs completed per tech per day, first-time fix rate, average ticket size through upsell prompts. Those three roll up into the big number - which turns "we help you grow revenue" from a vague claim into something you can actually trace.

As you go down that path, you can layer in firmographic, technogarphic and geo filters. The firmographics will be the ANSWER to the questions in SUcK.
- Multi-location home services companies,
- 20-plus technicians, meaningful revenue, running QuickBooks
Whatever the data and customer interviews support. These filters will actually work now, because you already know exactly what you're doing for whom, and why it matters to their career.

Notice what just happened. Firmographics showed up last, not first. That's the whole shift.

The ICP Becomes the Friggin’ Value Proposition

Run this properly and you land on a specific sentence:

We do [use case] for [champion persona] at [firmographic criteria], to help them improve [process-level KPIs], so they can nail [business-level KPI].

Sounds like the formula for a classic value proposition, right? For the field service example:

We do dispatch optimization for VPs of Operations at multi-location home services companies with 20+ technicians and meaningful revenue, to help them lift jobs-per-tech-per-day and first-time-fix rate, so they can nail their board-level KPI: revenue per technician per day.

I call this an artifact on purpose, not a pitch. Say it out loud at a dinner party and people take a slow step back. (Fair reaction, honestly.) That's not what it's for. It's the source of truth every downstream decision gets built on and tested against. Positioning, website copy, ad hooks, outbound, decks... When messaging drifts, and it will, (except when you use a GTM Brain) this is what it gets dragged back to. The test is always: are we still talking about the KPI, use cases in the situations that matter?

Here's why this beats firmographics-first. Firmographics-first gives you targeting - where to send the outreach. It has you guessing about where your solution matters. Firmographics should be the ANSWER to the S.Uc.K. parameters.

The SUcK process also builds 80% of your positioning, message and story.

If you're mid-ICP work right now, reply and tell me the use case you're building around - happy to sanity-check it. Or…

If your messaging doesn’t land » run this drill and see where it takes you.

Share

Take it easy:
Dan Renyi, founder at Klear

When AI got good enough to actually help with marketing, everyone in my feed asked the same question.

How many channels can I run now?

We asked the opposite one. What’s the one campaign we can pour everything into and make it bigger-better than ever before?

Subscribe now

Then we did exactly that. One GTM campaign. (A webinar-based motion, if you were wondering.)

80% of our effort on it, for months. The other 20% went to experiments and other marketing that we needed to do. We didn’t spin up a dozen experiments. We didn’t grab six new channels just because the tools suddenly made it possible.

(It was tempting though, I’ll tell you that.)

We watched the bar rise for everyone the second AI got cheap - and decided the winning move was to go deeper than anyone else was willing to, on one thing.

The Deal

Here's what you'll get out of the next few minutes:

• The case for why I made that bet, and proof it's more than a nice theory

• The real economics, including the numbers going the wrong way.

• A useful mindset shift: why should stop chasing efficiency and what to go for instead (this may be the most valuable part)

• Why I've stopped losing sleep over a rising cost per lead.

• And the one thing underneath all of it that no competitor can copy by buying the same AI tools you did.

If you're running more channels than you can love right now, and the results aren't matching the effort, this is the argument for pointing everything at one.

WTH is One GTM?

Quick reminder of what I mean by the One GTM Campaign. You pick a single motion - one audience, one message, one offer you actually believe in - and put the bulk of your firepower behind it for months, instead of spreading the same effort across every channel the tools just made easy to spin up.

It makes the most sense for most B2B tech companies that have just found product - go-to-market fit but don’t yet have the marketing and sales firepower to run several plays and campaigns and channels at once. When you can only fund one motion properly, funding one properly is the whole game.

Example of a One GTM campaign that can lift such a company to great heights: content»webinar»activation.

Or: content » signal tracking » ABM motion

The idea is to maximize the hell out of a motion that works until you can1t squeeze more out of it, before jumping on to a new one.

And it’s the power of focus, plain and simple: attention is the scarce resource now, not production. Concentrate it in one place and you get something with real weight behind it. Spread it everywhere and you get a pile of stuff nobody remembers.

Let me be honest with you about how this is actually going. Not the highlight reel, but the reality:

Our cost to acquire a customer is going up. Not down.

I run an agency. My workflows are wired tight - the automations, the research systems, the GTM brain that holds everything. And with all of that humming, getting great results still takes about the same effort it took three or five years ago.

That surprised me actually. I expected the machines to make it cheaper.

They didn’t. But they did make everyone using AI, including our competition, better. Which, if you sit with it for a second, is the same thing. Everyone got the same superpowers at the same time. So the floor came up under all of us. More content, better content, more outbound, more polished outbound. So it’s not just net noise, it didn’t just get louder. It got more competent.

Now, I told you we hit a $136 to $148 cost per qualified lead on that one campaign. Lowest I’ve seen on the European Saas market in while. Real number for genuinely qualified, sat meetings.

But let me not oversell it: that’s the exception, not the norm. Sorry. CPL is climbing for us too, campaign to campaign, same as it is for everybody. That one was a great run. I’m not going to pretend every Tuesday is like that.

And here’s the thing - I don’t lose sleep over any of it.

Because we’re not chasing efficiency. We’re chasing maximization. Those are two completely different games.

Efficiency asks the wrong question

Efficiency asks: how do I get my cost per lead down? Even “how do i do more with less?”. Wrong questions.

Maximization asks a better one: what’s the most I can profitably spend to get a customer - and how do I get as many of those as humanly (or AI-ly) possible at that price?

If a qualified sales call is worth $600 to you, the game isn’t getting the call for $200. The game is getting every single call you possibly can for anything up to $599. Then pouring everything you’ve got into filling that up.

Because at the end of the quarter, nobody pays my mortgage with a marketing efficiency metric. The bank account cares about one thing: how much profitable growth actually landed. Dollars in. Not campaign ROI% on a slide.

What would you rather have of these 2 campaigns:

A. put $100 and get $1000 profit, 10x ROI

B. put $1000 in and get $2000 in profit, 2x ROI.

(not gonna give you the answer :)

Once you see it that way, a rising CAC stops being a crisis. It’s just weather - like Europe heating up. Not happy about it but you dress for it and you keep maximizing.

Why “it can do so much” is a trap

So back to the dozen channels.

The instinct with AI is to spread. It can write the blog, the emails, the LinkedIn posts, the ad variations, the video scripts - all of it, all at once. So why wouldn’t you?

Because it can produce a lot. It can’t produce a lot that’s good enough to stand out. Not without a huge amount of human attention and love poured into each thing. And standing out is the whole job now. When everyone’s shipping decent AI content across every channel,

“Good content” now blends in with the noise. Spread your attention across twelve channels and you get twelve mediocre-okay presences. Concentrate it on one campaign and you get something with enough depth to actually cut through.

That’s the ONE GTM campaign. Not a limitation. A weapon.

What it looked like in practice

We had a retail tech client. One campaign. We built it with a ton of AI - including the videos. AI-generated video for a “book a discovery call” flow that pre-qualified leads before they ever hit the calendar.

But every piece of it pulled from real depth. The GTM brain held the ICP, the call intelligence, the competitor positioning, the actual language buyers used and (co-)produced many of the assets. So the AI wasn’t guessing. It was drawing from a well we’d spent real time filling and used workflows that weren’t invented that afternoon, but proven over time.

That’s how you get to $136 a lead. Focus, plus depth, plus AI aimed at one target instead of twelve.

Focus and depth. Not luck.

The same thing showed up where I didn’t expect it

Cold outbound.

The whole market is sprinting toward volume. More sending, more sequences, more AI-personalized-at-scale spray-and-pray. So we went the other way.

Low volume. Obsessive quality. Massive research behind the list, scoring, checks, re-checks, validation on every single name before anyone gets an email.

We book meetings at 2 to 2.5% of the target companies. Not of contacts emailed. Of target companies. In a channel everyone’s busy declaring dead, going narrow and deep is quietly working really well.

Everyone zigged to volume. We zagged to precision.

The part nobody can copy

Notice the thread running through all of this.

None of it works if you just stitch a bunch of AI tools together. You can wire fourteen apps into something that looks like a command center and still ship shallow garbage - because the tools don’t know anything. They don’t hold your context and memory.

The GTM brain does. It’s the backbone of the whole thing. Here is an aspect of what a GTM brain looks like (and you can read up on it here):

Subscribe now

And, yeah, the metaphorical brain too. Owned systems, buyer signals, market intel, call recordings, positioning - all of it in one place, feeding everything downstream. That’s where the depth comes from. That’s the part a competitor can’t get by buying the same stack you did.

The tools are a commodity. The brain is not.

Where I’ve landed

After pointing serious AI firepower at exactly one thing for months, here’s the conclusion.

The instinct AI creates - do more, everywhere, at once - is backwards. The cheaper it got to produce, the more crowded every channel became, and the more it pays to be the one who went deeper than everyone else on a single campaign.

CAC is going up. CPL is going up. I’m not fighting that. I’m making sure that whatever it costs, we squeeze the maximum profitable growth out of one campaign we actually believe in.

Small team? That’s not your disadvantage here. It’s the reason focus works so well for you. You were never going to win the volume war anyway. Good. Don’t try.

Pick the one. Point the brain at it. Maximize the hell out of it.

Share Dan Renyi's Electric B2B

If you’re running more channels than you can love right now, and the results aren’t matching the effort, hit reply and tell me what you’re seeing. I read every one.

Two pieces that go deeper on the machinery behind this:

• Content Waves - why I built a whole method around the fact that decent content doesn’t cut through anymore

• The GTM brain - what actually lives in the backbone, and why stitched-together tools never get there

Have a great weekend and happy 4th of July if you’re State-side.

Dan Renyi

The fastest way to find out whether a team’s marketing is mediocre/bad is to wire it up to AI and scale it.

You’ll hit the wall in a month instead of a year. When you’ve 10x-ed output on 7 channels and it moves nothing a couple months in, that’s a clear sign that FUNDAMENTALS are broken. So even a few days in, you might be shocked by the results:

That’s the half-joke: “scale it fast and if it sucked in the first place, at least you’ll know soon.” But you didn’t come here for that advice, so here’s what you’re getting:

How to build an AI brain and content mechanism that makes your GTM better instead of just “more”.

The thing you’re building - the GTM brain: your sales and marketing motions working and compounding the way a software product does. So, this construct - the AI system you want running your marketing - doesn’t fix what you feed it.

Subscribe now

It multiplies it. Good inputs get bigger. Bad inputs get bigger too, just faster. Most people building these systems are - knowingly or not - assuming the AI will improve the raw material. It won’t. It does exactly what you gave it to do, faithfully, at full volume.

So the real question was never which model or which tool. It’s what you’re feeding it. And that part has nothing to do with AI.

Look, I have multiple models hooked up to the same repo reading and working from my context:

Looks messy and confusing, right? Don’t worry, I’ll have up to 7 tabs open on a bad day.

Are you obsessing over the wrong layer?

Right now the whole conversation about AI in GTM is stuck at the tooling layer. Which agent. Which model. Which MCP. There are landscape maps, a shiny new “Agent Operator” or GTM Engineer job title, and an endless run of walkthroughs showing you how to wire fourteen tools into something that looks like a command center.

It’s genuinely interesting stuff. If you’re curious about it, fair enough - it’s moving fast and worth understanding.

But it’s also the layer that’s turning into commodity the fastest. The tools are converging. The models are converging. A year from now your competitor has the same stack you do, for about the same money. Whatever edge lives in this tooling layer is temporary and will be gone in 20 minutes (internet-time).

The GTM brain doesn’t break there. It breaks one layer up - at the fundamentals. Nobody’s filming a walkthrough of that layer, because it doesn’t demo well and it isn’t fun. It’s old-school marketing work. The stuff we were doing years/decades before any of this existed. It didn’t get less important when AI showed up. It got more.

Here’s where it breaks. At least in four places, but really it’s one idea wearing four faces: the system multiplies whatever you feed it, so what you feed it is almost everything.

1. Your ICP is a filter, not a strategy

The easy version of knowing your ICP is firmographic. Company size, industry, revenue band, tech stack. Add a few cheap signals like funding or hiring. You filter a list, you call it targeting. It’s easy because it’s the part a machine can do for you.

The real version is harder, and it’s manual. It’s your product’s actual capabilities, matched against the jobs you do better than anyone, matched against the companies and teams that have those jobs to be done. Not “B2B SaaS, 50-200 employees.”

More like

That work lives in your product, in what your team can actually pull off, and in your CRM - the real pattern of who closed and why. It does not live in a list you can buy.

I went 13 pages deep on this about a year ago. The one-line version: a GTM brain can’t find fit you never defined. Hand it a set of filters and it’ll faithfully bring you more companies that match the filters, and miss the point completely.

2. Connecting your CRM is not feeding your brain

This is the most common one I see, and it’s getting more common because the tooling turned it into a single click.

The instinct is: connect the CRM. Wire it in through an MCP server, give the brain the keys to everything, and now it’s “intelligent.”

Like,

Look ma’, it has all my data!

But connecting your CRM is not the same as feeding your brain.

Most CRMs are a mess. Duplicate records, dead deals nobody closed out, the same field filled in three different ways by five different people over four years. Pipe that straight in and you don’t get intelligence. You get a system that takes your mess and hands it back to you in confident, well-formatted sentences. Same garbage, it just sounds authoritative now, which makes it… sneaky?

The not-so quick fix

Build a database (in Supabase, e.g.) that contains data that we know is relatively reliable. If the CRM is a mess, skip the convenient MCP that pulls everything it finds into a Claude workflow and build a “mini but real” set of data. Yes, this can take the form of separate lists inside the CRM and you’d instruct the data pipeline to pull ONLY from the clean sources.

The context layer has to be curated. Filtered, known-good data. Use a direct CRM connection when - and only when - you actually trust the data is as clean as the job needs it to be. Otherwise you haven’t built intelligence. You’ve built a very fast way to be confidently wrong.

3. Mediocre in, mediocre at scale

Here’s where the joke from the top earns its keep.

If your positioning is mediocre - messaging that sounds like every other vendor in the category, a narrative anyone could have written - then doing it more efficiently doesn’t save you. It speeds up the problem. Now you’re producing forgettable stuff faster, across more channels, with more conviction.

The brain has no opinion of its own. No taste. The only point of view it’s got is the one you handed it. Weak point of view in, and you’ve built a machine for broadcasting weakness at scale. Mediocre in, mediocre out, just a lot more of it.

Now the genuinely useful bit, and I’m only half joking. Scaling something mediocre very fast has one real upside: it kills the mediocre version quickly. You see the flat results sooner, you’re forced to face them sooner, you get to rebuild sooner. I try to avoid this at all cost but can imagine a scenario as an in-house marketer who is being forced to do dumb shit in content.

Subscribe now

If there is no other way, I’d might just triple down on the dumb stuff to show the (lack of) results fast and bluntly.

But if you’d rather skip the dying-first part: fix the
1. ICP
2. positioning,
3. story and messaging
4. test and then scale it.
That’s the order.

4. More context is not better context

The last one is a belief that feels obviously true and is dead wrong: that more context is better context.

The pattern goes like this. People dump everything into folders in the repo. Every doc, every old deck, every transcript, every brand guideline they’ve ever written, on the theory that the AI will sift through it all and make the best of it. More material, more intelligence. Right?

Wrong. A brain you feed everything pays close attention to nothing. It can’t tell your sharpest, most current point of view from a positioning doc you forgot about eighteen months ago - sitting next to recent articles that reflect different value props.

They’re both sitting in your repo folder, equally weighted by the AI - it won’t know which one to lean into, and will not tell you that there is misalignment. You need to be the guard at the door.

The actual fixes are:
1. subtraction
2. autolearning
3. update loops

Step 1, the art and science of SUBTRACTION is with the human, you. Deciding what not to feed it. The quality of a GTM brain comes from what you chose to leave in.

Step 2, autolearning: installing memory beyond what Claude/Codex inherently have, and self-learning mechanisms. This is just rules and scripts that run after each session, that give our Claude-based brain the same memory-advantages that Hermes and OpenClaw have built in.

Another example: as the system does weekly media scan, it keeps track of what’s been talked about and can feed you the 2-3-month and 1-year trends…. that later feed into strategy and context documents.

Step 3, human-led update loops: this needs to be mostly human again, where you periodically tell the system:

Yes, this could be automated, but it’s one of the many places I like to have flesh-and-blood Homo sapiens keep control. You want to be aware of updates to your core context and strategy docs.

Have you noticed, how at least 2 out of the 3 meta-fixes to GTM brains come from humans and fundamentals.

The moat is the boring part nobody films

Line these up and they’re the same point from four angles. The system multiplies whatever you feed it, so the input work is massively important. ICP depth. Clean, curated data. A point of view worth having. The discipline to leave things out.

None of it is about AI, all of it is the old-school marketing and business fundamentals we’ve always had to get right. As in: even before the internet.

What changed is the price of getting them wrong

When execution was slow and expensive, sloppy fundamentals were survivable - you only ever produced a little bad work, so the damage stayed small. Now execution is instant and nearly free, which means your fundamentals get expressed at full volume the moment you flip the switch. The system doesn’t forgive you for skipping the hard part. You get a larger bill, faster. :)

The good news, and there is always good news: this is the part you already know how to do. It doesn’t need a new tool or a new model or a weekend course on agent orchestration. It needs the work you’ve been capable of for years, done honestly, before you scale it.

The tools get cheaper (well, not always) and more interchangeable by the week. The fundamentals are the moat. They always were.

Getting that input layer right - the ICP, the point of view, the context that actually belongs in the system - is most of what the first two weeks of a Content OS Sprint is. We do the boring, foundational part properly, so that when you scale it, you’re scaling something worth scaling.

So before you wire up another agent: what are you feeding it?

Hit reply and tell me what you’re building, where you are on the journey to build a GTM brain. I’m your biggest fan, rooting for you:

Dan Renyi

- Founder at Klear, a GTM advisory that now delivers results by installing a “GTM brain” your team can own&operate.

p.s.: I’m always worried that my articles push some of you into analysis-paralysis. Yeah, building an AI-based marketing or go-to-market system might be intimidating. Don’t let it stop you. Getting started and building intelligently is what matters. Just get going and iterate frequently. That’s how you build an advantage - and the path is very fulfilling. Do not obsess about getting every detail right - it’s impossible. Start small, go step by step and enjoy the journey.

This week’s digest made me put the coffee down. It took a moment to tie the pieces together, but…

Four opinions that are really one

Four people. Four corners of the market. No coordination between them that I’m aware of. All describing the same broken thing in B2B marketing - and not one of them using the same words for it.

1. Jay Acunzo said it in a single line: the brilliant thought leaders he follows have “basically devolved into unpaid journalists.” Everyone got the same AI tools at the same time, so everyone ships the same interchangeable stuff at the same cadence. Volume up. Authority down.

   

2. Dave Gerhardt put it thusly: brand is the only thing that matters now. 600 likes plus 492 comments agreed with him - way more than his usual. Talk about hitting a nerve.

   

3. Maja Voje shed light on a practical aspect of “most content is crap SO only brand matters” connection:
   
   14% of marketers feel confident in their AEO/GEO (=AI search) strategy. The other 86% are being researched by buyers inside ChatGPT right now, with no say in what the model says back. Not getting cited is half the problem. The model spitting back something wrong about you is the other half. Put more bluntly:

   AEO is becoming a very important discovery channel here to stay - and 86% of us are not doing jack about it.

   Fun experiment: Open ChatGPT and type [your brand] vs [competitor] differences and similarities. Many marketers freak out at what comes back and have no idea how to influence it.
   

4. Peep Laja spent the whole month dripping stats from Wynter’s own research. The stat that stuck: 53% of B2B marketing leaders say their AI investment hasn’t paid off. Not because the tools are weak.

   Because they aimed them at the same broken process and ran it faster.

   

   
   AI is even more dangerous than coffee, ya know:

   

   Subscribe now

Do you see the common thread?

That’s an actual question. Go back for 15 seconds and re-read those four observataions, before proceeding. Do you see the bigger story they converge on?

None of them set out to make the same point. That’s what got me. But four pretty smart people with nothing in common described the same beast from four sides.

Here it is.

The numbers aren’t kind. Ahrefs looked at a billion pages: 96.55% get zero traffic from Google.

And the pile keeps growing - 74% of new web content now has AI in it. More words, less of it read, every day.

Now look at the team down the hall. Product ships every sprint, and every release stands on the last one. That compounds the way you get compounding interest on a financial investment.

Compare that to marketing:

• New head of marketing? New direction, new strategy. Scrap what we had.

• New agency, discovery from scratch and reinventing a slightly different wheel.

• Reorg? Four months of messaging work, gone.

What nobody says out loud

Publishing content and building demand are two different jobs. Sadly, there is weak cause-effect or correlation between the two. Most companies do the first one and pray it turns into the second.

They

1. get someone to do content and

2. push it on a couple of organic channels,

3. maybe put some paid distribution behind it.

4. have zero clue why more inbound leads don’t happen.

If that’s you, doing that treadmill because it worked at your previous job five years ago, don’t feel bad, but people are waking up to this reality.

The factory - brief in, post out, tweet-share, next - does keeps everyone busy but moves nothing.

Buyers still don’t know you. And they’re deciding without you: 94% of B2B buyers used an LLM during their last software purchase (6sense). If the LLM doesn’t know who you are, you’re not in the room.

So I spent two months building the opposite of a factory.

Content as a product. A real lifecycle - research, create, launch, distribute, measure, relaunch - the same way your product team treats a feature.

The thing that cracked it open:

The unit of work isn’t the content piece. It’s the Wave:

• Swell: two weeks of teasing before anything drops.

• Break: the launch, every channel at once.

• Wash: distribution, follow-ups, and a relaunch at 30 and 90 days.

Definition time: the Content Wave is

1. taking multiple pieces of content,

2. get them to work together to get attention, make an impact, and

3. have long-lasting, far-reaching repercussions.

Your team only focuses on Breaks. Most teams halfquarter-ass teaser-buildups (the Swell). They post on launch day, repost twice, and that’s it. Then wonder why it’s dead in 3 days.

“Why ain’t my content building demand?”

It’s actually not super easy to build a Wave - but it’s simple enough. That is, if you use the seven-part diagnostic. Honestly, it’s the basics. Again, not easy; 5 fixes in this order:

1. fundamentals - ICP, buyer personas, jobs-to-be-done, messaging etc.

2. narrative - do you have a story that pulls in the ideal buyer?

3. depth - is your content deep, meaningful… share-worthy?

4. clarity - is your content digestable, convenient, enjoyable, heaven forbid: fun? …share-worthy?

5. quality AND consistency - Are you showing up with great stuff regularly on the channels that matter?
   

   — above is the diagnosis - the building blocks —

   
   The ingredients needed but not enough for breakthrough reach, attention, demand creation. You need two other components:

6. stacking - are you building Waves for important content?

7. distribution - do you have content/media partners, employees that help distribute?

I’ve been experimenting with this in public, sharing progress and results in two live workshops. The funny part is I didn’t engineer this to be a Wave or even a series.

Wrote one question, chased it to the next, and twelve posts later they’d stacked into an argument people started resonating with, picking at, advising on...

• The “reset problem” made the content-as-product thinking necessary - spotted it on Jason Gong’s Li feed and it really stuck.

• The “nobody reads our stuff” problem made Waves necessary. Also blends perfectly with Content-as-Product.

• The “OK, but Waves don’t work” problem needed the diagnostic - because a method/playbook/framework always needs solid foundations.

  
  Each piece earned the next.

What everything points to

So: volume up, authority flat, half the AI budget wasted. Four strangers, one diagnosis. The market is arriving here on its own: Good content on its own does not build demand.

I’ve shown you the problem and the fix:

1. treat content like a product, with a lifecycle,

2. with Waves as the unit of optimization,

3. the seven components in order.

What I haven’t shown you yet is what it looks like to build the whole thing for one company, start to finish, in 18 days. Because, this thing needed to be tested out in the wild.

My media brand, Electric B2B with my Linkedin presence was the perfect guinea pig. We need to make a big splash with one founder marketer who can’t commit more than 40% of his time on content.

Results looked pretty darn convincing, so it was time to take it and see if anyone is willing to pay for this AND get results.

Nice Concept, But Does It Work??

I’ve tested it on two clients so far - a highly irresponsible founder and a similarly risk-tolerant marketing leader. Clearly gamblers with issues. Both paid for a full Content as Product buildout and a couple of Content Waves.

The hard numbers take another few months to land. But they’re loving what they’re seeing. The output and the process.

For the next 3 equally reckless GTM leaders who need to build demand, show up in LLMs, and actually produce pipeline - not just publish - Klear has founding member pricing for a limited time.

Is Klear’s track record any indicator of future success? Honestly, not sure. If yes, these brave women and men will be very pleased. Why? Take a look at a couple demand gen results:

We’ve got more testimonials on this page and case studies on the website.

I need to build demand NOW

To be fair: the first three of those results are pre-AI.

• Same processes,

• same knowledge

  just without the system I've since built on top of it. The Content as Product buildout combines 12+ years of B2B SaaS/tech content-based GTM experience. Six of those years with Klear and its clients.

That’s it. That’s the pitch. No “this offer expires in 47 minutes.” Just three spots, founding pricing, and a guy who’s been doing this for 12 years and is genuinely curious whether his clients will become case studies or cautionary tales.

(They’ll be case studies…………………………………………… Probably.)

Reply or hit one of the many CTA links above.

- Dan Renyi

ps.: I published a big resource on this topic with one of the training recordings, slides and more. Sent in a newsletter at 6pm last Friday (no idea why I hit send at the worst time of the week.) So lots of you missed it but here it is again. I’ll be taking the resource page down next week, as I go in Content Wave build mode with some of you.

pps.: Because I am a former PhD dropout, I take citations seriously. If you wanted to check out the sources for the stats I was throwing around, here:

- “96.55% of pages get zero traffic from Google” — https://ahrefs.com/blog/search-traffic-study/

- “74% of new web content contains AI-generated content” — https://ahrefs.com/blog/what-percentage-of-new-content-is-ai-generated/

- “94% of B2B buyers used an LLM during a software purchase” — https://6sense.com/resource/b2b-buyer-experience-report/ (6sense, cited in Forrester 2026)

- “Only 14% of marketers feel confident in their AEO strategy” — Maja Voje LinkedIn

- “53% of B2B marketing leaders say their AI investment hasn’t paid off” — Peep Laja LinkedIn feed

I put together a resource on the GTM Operating System and made it available for 7 days. 3,186 GTM people accessed it. The feedback was very encouraging. (So was the pipeline tbh.)

So we’re doing it again.

Once a quarter, one topic, full depth, limited time. Q2’s resource just went live - and this one’s about probably the most important pillar of GTM: demand generation.

Get it. Like, now.

The half-typed post

A few years ago I was halfway through a LinkedIn post in a Budapest cafe when I stopped. Not because I got stuck - because I realized this one was going to land, and I’d had enough good posts disappear before; 100 likes, 40+ comments - gone and mostly forgotten in 72 hours. I wasn’t doing that again.

So before I hit publish, I set up a webinar and ended the post with a sign-up link (in the comments, of course, where else!?)

My prevoius record for a webinar was less than 30 registrants. But to push this webinar even more, I added another mechanic with a self-assessment.

This webinar got 116 registrants.

For the first time, the thing ended with real pipeline. Not just engagement. Pipeline.

That experiment showed me something most content teams spend years not figuring out.

The unit of optimization in B2B content isn’t a single piece of content - not an article, not a post, not even a webinar. It’s something bigger. I’ll show you what in the resource.

What I’ll say here: most pieces of content are parts of larger Waves, not just assets you publish and hope stick. And once you start running it that way, the results are a bit hard to ignore.

Just 2 real,

flesh-and-blood examples from Klear’s past:

1. Running Content Waves: Infinityn - the kind of company with every reason to sound like every other vendor in a crowded, mature category - tripled in size and hit #91 on the FT1000 fastest-growing companies in Europe.

2. An e-com automation client used a Content Wave play to set up a product relaunch: pulled 41% MRR growth. From.A.Single.Motion. ContentWaves, ladies and gents. Not luck.

Then AI Happened

ContentWaves with AI is a different animal. Genuinely. AI has been the lever to get Klear’s Content Waves off the ground as the sole marketer at the company. I only have 39% of my time (clocked it for 3 weeks for real) to spend on content-stuff. This means my business only has 0.39 content people.

So 2024-25 was a mega jump. This newsletter went from 550 to 7,500+ readers. Forbes Hungary ranked me fifth on their most authentic LinkedIn creators list in 2025. Not first. 5th. (I’ll take it.) Every clean run of the Wave pulls 100-300K+ in pipeline. Brands now pay to sponsor the channel.

None of that happened from posting more.

You’re Getting the Content Waves Mega-Resource

The full May 26 webinar recording. The slides. Everything I know about building demand that doesn’t reset every month - including things I’ve never written down this clearly.

Specifically:

• The single biggest reason content stops working at month 12 - and the fix most people never look for

• Why “better content” is the same problem with a haircut - and what actually moves the needle

• Why Infinityn went from generic to category-defining (had nothing to do with writing better content)

• The AI layer most demand gen strategies are still missing in 2026 - the one that determines whether buyers find you or your competitor during the research phase you can’t see

• The maturity model that shows exactly where your content operation sits - and what separates the teams building moats from the ones starting from zero every week

Like in March, this is limited time. Last time it was 7 days, 3,186 people.

Access the Q2 Resource - ContentWaves:

Access it.

If you want this built rather than studied: the Content as Product Sprint. 18 days, full system installed, first Wave ready to run: super early-bird pricing until June 7 - click for details.

More soon.
Dan Renyi
- Electric B2B author, hobby tomato-grower, hamster owner, Founder at Klear B2B

ps.: The replies over the past few months have shifted since we started talking about AI GTM and content systems publicly. Less “can you do this for us?”

More “can you teach us how to do this ourselves?” I’ve been resisting but it’s clearly a demand, so still figuring out the right format to do this. This list hears it first, if it does happen.

BUT if you’d rather build than buy - reply to this or DM me.

pps.: if you stayed till the end of the live webinar: promised to send a couple of Claude files you can copy into your repo, so that your content operating system will have a brain - it’ll remember things and learn. You’re getting that in a separate email.

“We do a bunch of content, but not much happens.”

Guys, if I had a dollar for every time I heard a founder/GTM leader say this, I wouldn’t need to do content anymore. I’d just buy an island.

It’s a bloody frustrating place to be. You’re putting in the work. Content cal full in Asana, process grinding, Slack notif’s pinging. You’re shipping blogs, newsletters, and LinkedIn posts. Team reshares your work.

But when you look at the i.m.p.a.c.t.? Crickets. The brand feels flat. The “needle” isn’t moving or if it is, nobody thinks content has a lot to do with it.

“Just do more …and new!”

Most people look at this situation and think, ”We need to try a new channel” or ”We need to post more often.” They sprint toward a new tactic.

But tactics won’t fix a broken system. If your content isn’t working, it means there’s a leak somewhere in your engine.

Over the last few years, we’ve audited dozens of B2B content programs. We found that the ones that fail always break down in one of seven specific areas. And the order in which you fix them matters immensely.

If you fix step 6 but step 1 is broken, you’re just distributing noise.

Here is the 7-step diagnostic.

Where is your leak?

Share

The components and systems 1&2 explained 👇🏻

0. Fundamentals

You can’t build a house on quicksand. Who exactly are you writing for? What is the bleeding-neck problem you are solving for them? What is your core positioning? If these answers are blurry, your content will be blurry.

Today, blurry content = content forgotten faster than you can say “thought leadership”.

When you try to speak to everyone, you speak to no one.

This context/research layer - now living in a repository and used by AI agents - constantly learns, researches and updates the workflows.

1. Narrative & Point-of-View

Are you leading a conversation, or just echoing what everyone else is saying? Your POV is your ultimate moat. “How-to” guides get commoditized.

A distinct narrative - a strong belief about where your industry is getting it wrong and how to get it right - cannot be copied. Without a POV, you’re just another vendor. With one, you become a thought leader.

2. Depth and Consumability

Are you going deep enough to actually change a reader’s mind? Surface-level fluff is ignored. But depth alone isn’t enough - it has to be consumable. High signal, low friction.

Can your buyer absorb your deepest insights while drinking their morning coffee?

If it reads like an academic paper, you’ve lost them.

3. Clarity and Engaging

B2B does should not stand for Boring-to-Boring. If your writing is dense, corporate, or full of jargon, people will click away. You need sharp hooks. Write like you speak. You need to respect the reader’s time by getting to the point. All the while keeping ‘em engaged with rhythm and pacing.

Expanding on the morning coffee example from Component 2:

Will your writing make them choke on their coffee, drop everything, and sprint to their keyboard to act?

It’s not just about insights (value) but delivery, style, connection.

4. Quality and Consistency

You can’t just be brilliant once a quarter. You have to show up consistently with high quality. Consistency builds trust; quality earns attention. Most teams sacrifice one for the other—they either post garbage every day, or a masterpiece once a year. The winners find a way to do both.

5. Content Stacking

Most of content is isolated, one-off pieces. What a waste. Content can act like chapters in a book, teasers for a movie, awesome fireworks that draw people in and disappear...

Example:

1. Piece A introduces the problem and leaves a curiosity hook - teases Piece B.

2. B unpacks the solution but sets a few honey traps. “Can’t wait for Asset C to drop!”

3. Asset C builds anticipation further with a couple case studies… a self-assessment…

4. Piece D drives it home. It’s the solution to the problem. Piece D can be an event, a big report, even a product/feature launch that provides a solution to the itch you introduced throughout.

Each piece earns the next. When content stacks, attention compounds. Without it, the AI-slop-noise buries your valuable content.

Oh, and in case it isn’t clear - you don’t need attention for attention’s sake. You’re not an influencer. The buyer’s attention is a MUST, a non-negotiable if you want her to buy.

6. Distribution

Publishing is not distribution. Hitting “post” is the starting line, not the finish line. If you don’t have a system to amplify your work—creating the Swell before it drops, the Break when it launches, and the Wash in the months after—your best work will die in silence.

The hard truth and the good news

Looking at that list, you might be thinking: ”Great. Now I ‘only’ have to fix seven extremely difficult things. Thanks, Dan.”

It feels overwhelming. It feels like you need a team of ten people to get it right.

But fear not. You really can do this. You can build a content engine that drives real revenue, and you don’t need a massive team to do it.

My advisory firm, Klear has approximately 0.39 content marketers. That’s actually me, in our small company, spending 15 hours a week on average on content-related stuff.

And we freakin’ check all these 7 boxes. I want to massively improve in a few components, but hey, it’s never finished, is it?

Subscribe now

The secret isn’t working harder, but having a self-learning, smarter system.

We’ve built a system that takes human ingenuity; your unique POV, your expertise, your taste… and makes it work consistently and at scale by pairing it with AI.

It has two parts:

System 1: The ‘Content Product’ Layer.

In my house, content is a product, as I’ve explained earlier. It’s the

• self-learning

• auto-researching

• self-updating

context-research-production layer. Gemini and Claude agents and workflows. AI is only as good as the context you give it. We’ve built a layer that captures your fundamental truths, your narrative, and your specific tone, and creates things. The system also stops the AI from doing dumbshit things and turns it into an extension of your best subject-matter expertise and thinking.

System 2: The Content Direction Layer: The Content Wave.

This is the attention-orchestration component. This system takes the raw materials from System One and structures them into stacked, consumable assets ready for distribution. It plans the Swell, the Break, and the Wash, ensuring your content actually reaches the people who need it.

When you put these two systems together, content stops being a linear assembly line that nobody really enjoys and turns into a predictable, compounding asset - a PRODUCT for the business.

Now for the good great news: next Tuesday,

we’re showing you exactly how to build the two systems

We’re hosting a live workshop where we will pull back the curtain on this exact 2-part system. We’ll show you how to plug your own expertise into the context layer, and how to use the direction layer to orchestrate your next Content Wave.

You’ll walk away knowing exactly how to fix the leaks in your 7 areas, without burning yourself out.

Spots are limited because we want to keep it interactive. Get in on this:

Stop settling for content where “not much happens.” It’s not fun and certainly is not profitable. Build a system that is both.

See you Tuesday?

Dan Renyi

Hey Everyone,

Here’s a pattern I recognize right away in almost every B2B content engine I look inside; because I used to run demand like this:

1. A blog post goes up.

2. A couple people from team link to it/reshare and hit “like”

3. The newsletter mentions the blog and the upcoming webinar.

4. The webinar runs. A follow-up newsletter links to the recording.

5. Everyone moves on to the next piece.

That looks like a content engine. Hell, it even feels like a content engine: the calendar is full, the Slack channel is busy, and the dashboard shows “content published this month: 12.”

But the pipeline is flat. The brand still feels generic. Sales reps still don’t have anything they’re proud to send. And six months in, nobody on the team can describe what the company has been saying for the last quarter.

You’ve connected the pipes. You haven’t made the water move.

Why more or better pieces won’t fix it

This is linear publishing. It’s the default mode for most B2B teams.

The shape is always the same: one asset publishes, a few other assets link to it, everything goes live in roughly the same week, and then it dies. Each piece sits on its own little island, with a few hyperlinks running between the islands.

The hope is that someone will hop from one to the next.

Subscribe now

Almost nobody does.

The reason isn’t quality. Plenty of teams write good content. The reason goes deeper.

When you publish a single piece into a feed full of competing noise, you have a few hours of visibility before the algorithm moves on. If you didn’t seed anticipation before, didn’t build for amplification during, and didn’t relaunch after - that piece has done its entire job in less than a day.

The sad truth: today even great, kickass content needs to earn the attention it deserves. Greatness in itself is usually insufficient for reach and impact.

You can stack twelve months of these on top of each other and end up with what you started with: twelve months of small ripples. They don’t add up to anything meaningful. No prospect mentions content on sales calls and the stuff remains on your website like a stack of receipts.

This is quite a shitty feeling as a marketing leader (ask me how I know) because content is the most visible thing you’re judged on.

You conclude you need more content. So you hire someone, contract an agency, or buy a tool; or even pimp up your Claude content workflow. The new output gets added to the same calendar and the result is the same f*cking flat water, where nothing moves. It lacks energy.

More pieces won’t fix this. The unit of work is wrong.

Swell, Break, Wash: the anatomy of a wave

Here’s a different way to think about generating demand with content: it’s not about the sum of individual pieces.

The new unit of measurement: the Wave. A wave is multiple content pieces working together to maximize reach and impact.

A Wave isn’t a single blockbuster viral piece of content. It’s a system. And it has three phases.

Share

Swell. The build-up. The 3-10 days before your main asset goes live. You’re seeding curiosity, teasing the question, planting the contrarian data, dropping hints in your network. The audience starts leaning forward before the asset drops.

Break. The launch moment. Not just publishing the piece. Hitting from every direction at once - LinkedIn post, newsletter, partner co-promotion, sales rep outreach, paid amplification if you do that. The asset arrival feels like an event - not as a publication.

Wash. The days, weeks, and months after the break. The piece keeps reaching people who missed it.

• New hooks.

• Repurposed angles.

• Refreshed data.

• Quotes added.

The asset has a long tail because you built one - intentionally.

Most teams skip Swell and Wash entirely. They show up on Break-day with a fresh asset and expect consumption and attention to happen automatically.

They don’t. This hope is the content equivalent of “build it and they’ll come.”

What creates the wind

Flat water doesn’t become waves on its own. Something has to move the air above it. In content, that something is a set of practices that compound each piece’s reach and impact.

Five of them. Together, they’re the wind.

1. Narrative arc

Your pieces are chapters of one story. They aren’t standalone articles that happen to live on the same blog.

What this looks like in practice: piece one introduces a problem and hints at a counterintuitive answer. Piece two unpacks the counterintuitive answer with data. Piece three shows what happens when companies act on it. Piece four answers the objection most people will raise after reading piece three.

Each piece is satisfying on its own. But each one also creates pressure to read the next. The reader feels walked through a thesis, not handed a folder of disconnected articles.

If you can’t describe your last six months of content as a coherent story with tension that builds and a payoff that lands, you don’t have a narrative arc. You have an editorial calendar.

The fix isn’t to publish less. The fix is to plan the story before you plan the calendar.

2. Content stacking

Build assets that depend on each other.

The article ends with a question it deliberately doesn’t answer. The webinar answers it. The follow-up LinkedIn post unpacks what the webinar revealed. The case study illustrates what happens when someone applies the answer in practice.

Each piece earns the next.

This is the opposite of what most teams do. Most teams write self-contained pieces. Every article tries to be the definitive take. Every webinar tries to cover the whole topic. The pieces compete with each other for the reader’s attention instead of cooperating.

Practical move: before you write the next blog post, decide what asset it sets up. If it doesn’t set up anything, it’s a one-off. One-offs go flat.

3. Teasing (the Swell phase, made deliberate)

Launch a content asset the way you’d launch a product.

Two to four weeks of Swell before the main piece drops. This is the part teams skip almost universally.

What Swell looks like in practice:

• A LinkedIn post that names the question the upcoming piece answers, without giving away the answer. “We just finished a teardown of 30 B2B content engines. The patterns are nothing like the playbook says. More on Tuesday.”

Behind-the-scenes glimpses. The research you did. The unexpected thing you found. The hypothesis you started with vs. what you ended up believing. All hinting at the reveal that is around the corner…

• A direct ask to your sales team and channel partners to share the upcoming piece when it lands. Not asking once. Asking three times, in different ways, with the actual links and copy ready to paste.

By launch day, the audience is leaning in. They know the piece is coming. They know what it’s about. They know they HAVE TO pay attention.

Compare that to the default: piece goes live, link goes out, hope works.

Product launches don’t operate that way. They have pre-orders, beta lists, teaser videos, founder posts. Content can borrow the entire mechanism.

4. Content partnerships

Co-produce with someone whose audience overlaps yours.

One webinar, two newsletters, two LinkedIn networks, two sets of sales conversations. You haven’t doubled your reach. You’ve multiplied it, because the partner’s audience trusts the partner in a way they don’t trust you yet.

Subscribe now

The partner doesn’t have to be famous. They have to have an audience of your buyers.

A category-adjacent SaaS company. An advisor who works with your ICP. A community founder. A journalist who covers your space. A consultant whose retainer clients are exactly your prospects.

This is the cheapest distribution layer most teams never activate. The reason isn’t budget. The reason is that it requires creating distributable assets deliberately, which most teams don’t, because they don’t think of content as something other people would actively want to share.

5. Content boomerangs

This isn’t (just) repurposing, it goes further.

Thirty days after launch, take the same piece, write a new hook, and send it to the people who missed it the first time. Most of your audience didn’t see Wave One. (Even if they did, they probably don’t remember.) They’ll happily read it now.

Ninety days after launch, refresh the data, add the new angle the conversation has moved toward, and ship it again.

Six months later, when the topic comes back into the news cycle, reference it. Quote yourself. Build the next piece on top of it.

A good piece of content has at least three lives. Most companies give it a half.

This isn’t rocket science, but it does require 1. imagination, 2. discipline. Teams want to ship new things, not relaunch old ones. The new thing feels productive. The relaunch feels lazy.

Not true. Think of relaunches as squeezing more ROI out of a past content investment. The relaunch is where the compounding happens.

Same effort. Different shape.

Five ingredients:

1. Narrative arc,

2. content stacking,

3. teasing,

4. partnerships,

5. boomerangs.

None of them are exotic. None of them require new tools and hey, most of them don’t even require new content! They require thinking about your existing output as a system instead of a sequence. Think waves instead of ripples.

A team that adopts even two or three of these might even produce fewer pieces, but with more impact. Same effort, different shape. Waves instead of lines.

The default is a flat ocean with a few hyperlinks between rocks. The alternative is Maui

Workshop: map your first wave

If you want to see what this looks like applied to your own content - turned into a six-month plan with the Swell, Break, and Wash phases mapped to specific pieces and channels - I’m running a workshop on this.

It’s hands-on. We map your next wave together. By the end, you’ll have a plan, not a theory.

Reply to this email with “WAVES” and I’ll send you the details.

- Dan Renyi, founder at the agency that’s been using Content Waves to build demand; even before I came up with the term.

Put that out on Linkedin this week (with some elaboration) and the comments lit up. Must’ve hit close to home.

• the founder who’s been writing posts no one shares,

• the CMO whose content factory is humming and not moving pipeline,

• the marketing leader a year into something that feels kinda off……… but can’t quite name why.

That kind of response tells you the conversation is ready to happen out loud.

Which is why we’re doing it live and bringing you the current state of the (ever evolving) solution. May 12. Together.

Before we get there, quick recap of where the last few weeks have landed.

The number that started this whole thing

Earlier this year, Dreamdata reran their B2B buyer journey research. The number came back at 80%.

Roughly 80% of the B2B buying cycle now happens before the buyer ever talks to a salesperson. Up from the 70% most teams were already operating against.

Of that self-serving 80%? Around 90% runs through content. Articles, newsletters, videos, podcasts, comparison pages, frameworks. Whatever the buyer can find on their own time and form an opinion from.

Translation: marketing is no longer the demand-gen accessory that hands leads to sales. Marketing is now responsible for whether buyers self-educate their way toward you, or away from you.

By the time they book a call, most of the decision is already made.

This isn’t subtle. It changes what content needs to do, who it has to reach, and how good it has to be to count.

Why “build a content engine” is bad advice

Once you clock that 80% number, the obvious response is to crank out more.

Build an engine. Build a factory. Add posts. Add articles. Add videos. Add newsletters. And while you’re at it, make ‘em “more engaging.”

That’s been the dominant move across B2B for the last 18 months. And it’s where the most expensive mistakes are getting made right now.

The factory optimizes for output. Cadence. Quantity. How many pieces went out this quarter. How many touches per account. How many words got drafted, edited, scheduled.

It does not optimize for impact.

Volume gets reported in dashboards. Impact shows up somewhere else: in pipeline, in sales conversations that start from a different place, in buyers self-qualifying before they ever fill out a form.

The factory model can scale the first one infinitely while making the second one quietly worse.

Smart, well-funded teams are stuck right inside this trap. (Fair, honestly. The instinct to make more content is the correct instinct.) The execution model is what’s broken.

Think like a product team, not a publishing team

Here’s the reframe that’s been resonating.

Think about how a software product gets built.

Every feature is designed to create impact for the user. Every update makes the product better. Every release teaches the team something they fold back into the next one. That’s what keeps people coming back. That’s what drives retention.

Your content should work the same way.

What you actually need is not “more posts.” It’s quantity AND quality AND something qualitatively different. Content that compounds. Content that earns the right to exist. Content that pulls people back.

That’s a different operating model. Not a faster factory. Not a tighter calendar.

A product lifecycle: research, draft, iterate, launch, distribute, measure, relaunch. Each cycle leaves something structural behind. The system gets smarter every month, not just busier.

There’s a progression worth naming, even briefly:

Calendar. Deadline-driven. Volume-oriented. Manufacturing mindset.

Flywheel. Cycle-driven. Iteration-oriented. Momentum mindset.

Infrastructure. Compounding-oriented. Scale mindset. Every cycle leaves behind something structural that makes the next one easier and sharper.

Most of B2B is stuck on the calendar. Some teams have made the leap to flywheel. Almost no one has made the leap to infrastructure - which is exactly where the real compounding lives.

The honest question we keep coming back to is the one that proven LinkedIn post left wide open:

How do you actually design content that makes a BANG? Content that moves buyers? Content that earns its place in their week?

We’ve got answers we’re testing. Wins we want to show. And things we’re still working out in real time.

That’s why this is live, not recorded.

Come work this out with us, live, on May 12

This is where everything lands in the same room.

The 80% number and what it actually means for marketing teams under pressure right now. The factory trap, and why so many serious operators got caught in it. The shift to content as a product. The maturity ladder. What we’re seeing work. What we’re still iterating on. What you can take into Monday morning and start operating differently.

What you’ll walk away with:

A sharper read on which stage your content operation is actually at, and what the leap to the next one takes.

The product-team approach to content, broken down into something you can actually run.

Specific patterns we’re seeing compound. And a few we tried that didn’t.

Live Q&A. Bring the version of this problem you’re actually dealing with.

[Date / time / format placeholder] [Registration link placeholder]

If any of the last few weeks landed for you - the posts, the newsletter, the comment threads - this is the room to be in.

Save your seat. Bring your questions.

See you on the 12th.

Earlier this year we compiled five buyer journey studies into a meta buyer-journey report. They all kept circling the same uncomfortable number:

Roughly 70% of the B2B buying cycle now happens before the buyer ever talks to a salesperson.

Then Dreamdata re-ran their version of the research. The number moved.

It’s now 80%.

(Sit with that number for a moment.)

That isn’t a rounding error, right? It’s a deep/fundamental/grassroot-level change in how B2B buying works. And it’s not about to reverse.

80% of the deal is decided (mostly) before your sales rep gets to say a word.

The game has changed. Your buyers are now doing what you are, when you’re shopping for a B2B solution:

The discomforting small print: That graph is from late ‘25. Every datapoint since then suggests that AI vetting-shortlisting-screening is even higher and rising.

This is how important LLMs have become in the buying journey, relative to other sources:

So:

Winning the deal isn’t primarily about how your account executive handles the first call anymore. It’s about whether you won the 80% that happened before the first call.

That shifts enormous weight onto marketing. Enormous.

Roughly 90% of that self-serving happens through content in one form or another - articles, videos, podcasts, tools, comparisons, frameworks, newsletters, posts, webinars.

So reading this, you might be like

“do we just build more and better content? Isn’t everyone trying to do that already?”

the answer to both questions is yes, but it requires a different way of thinking; and certainly a different way of producing content, because

the “content factory” you have built is already bleeding.

(Spotted: an “AI-powered B2B content factory”)

Once founders and marketers clock that the majority of the journey now rides on what buyers can find on their own, the instinct is almost always the same.

Crank out more stuff. Build an “engine.” Build a “factory.”

Quality-questions and AI-slop aside, we have a deeper problem: the factory model is a one-way assembly line.

You give the machine input in the front. A finished post drops out the back. Somebody edits it, somebody schedules it. Next week, repeat.

This treats content as a linear process that it is not it shouldn’t be. The goal becomes output and some vanity metric slapped on to it.
- How many posts per week?
- How many views? How many articles per quarter?
- How many videos…….
can we ship before someone asks why pipeline isn’t growing?

Here is a mindset shift that will help transition to not just more content but more impactful content.

Content is a product, not a post

If the factory model is broken, the alternative is the product model. Product, as in how a software product gets built. Similar to how we talked about GTM (go-to-market) needing to become a product.

I’m going to break down the process for you. While we’re at it, what do you say I give you an agent I use that’ll execute the process below?

If you have Claude Code (or Gemini CLI), you can use this skill... an agent (and subagents) that will walk you through brainstorming, iterating and distributing kick-ass pillar content. Looping you in every step of the way.

What to do:
You can clone my git repo by pasting the following command in your terminal:

git clone https://github.com/zsalti/pillarcontentcreator.git

Subscribe now

That should install this skill for you. While you’re at it, consider giving it a star in the repo.

Here is the walk-through of how the above skill/agent works:

1. Braindump or briefing: the rough outline, talking points, and the main goal of the content piece.

   1. as you’re providing the braindump, a sub agent searches for context: Strategic inputs. Reference folders and files on your ICP, buyer personas. Reference product marketing docs. You need to have these in your repo in advance but the sub-agent walk you through uploading them.

2. Research agent: Go deep on the topic from multiple angles and get quotable data stats from verified sources. Make sure you use trustworthy sources, not vendor websites and claims in case studies. After you’re done researching, launch a sub-agent that will verify every single claim your research has found

3. Find the right angle. Provide 3-4 angles for the content piece that we want to focus on.

4. Content creation. You draft. But the draft is not the deliverable. It’s a prototype.

5. Iteration and refinement. You run it through feedback loops. AI and human. Stress-test the argument. Sharpen the messaging. Cut what’s soft. Refine until the piece earns the right to exist.

6. Select hooks: brainstorm hooks.

7. Apply brand voice. Rewrite in the brand’s tone of voice and apply content rules.

8. Launch. You don’t just “hit publish.” You launch - the way a product team launches a release. Real effort goes behind it.
   

9. Distribution. This is not an afterthought. It’s an integral part of the content-product. A great asset distributed casually will lose every single time to a mediocre asset distributed with intent. Channels, formats, repurposing, amplification, sequencing - all planned up front, as part of the asset itself.
   
   There are three main parts to distributing:

• pre-launch teasing: mostly on social media

• launch: distribution channels such as newsletter, website, video platforms, social snippets, partners, employees…

• re-launch: whenever the content gets updated or just because enough time has passed, the content can be relaunched - redistributed. See next point:

11. Measurement, relaunch. Once it’s live, the process doesn’t end. Good content, like good software, gets version updates. You watch how it performs. You listen to the market. You mix in new data, new sharpness, new examples. Then you relaunch.

That’s a product lifecycle. Not a publishing schedule.

Share

Calendar, flywheel, infrastructure - pick one

1. The primary mechanism companies have used to manage the content factory is the content calendar. It’s the backbone of even advanced content production systems like the one above.

But you don’t want to stop there.

Plenty of content teams have already moved on to a flywheel approach, and that’s a real step forward.

2. The flywheel recognizes that content and marketing don’t happen in one-shot deadlines. They happen in cycles. Launch, learn, iterate, relaunch. That is how you build on previous momentum.
   
   (Shout out to everyone who already made this shift. Seriously.)
   

3. Content-as-product. What we’re building inside Claude Code is the next step beyond the flywheel.

It takes the flywheel and turns it into infrastructure by building in versioning and self-improving mechanism.

Here’s the progression worth noticing:

• Calendar: deadline-driven, volume-oriented, manufacturing mindset.

• Flywheel: cycle-driven, iteration-oriented, momentum mindset.

• Product: infrastructure-driven, compounding-oriented, scale mindset.

When the flywheel becomes infrastructure, every cycle leaves behind something structural. A sharper narrative. A versioned asset. A reusable research base. A documented learning. A repeatable distribution playbook.

Get Another Brain

A “Brain” built in a repo gives you a living home for up-to-date context, research, rules, guidelines, automations. It holds your strategic briefs. It tracks performance. It stores learnings where both AI and humans can actually find them and use them.

That impacts quality in a huge way. Better content across the board and not because someone worked harder this week - because the system got smarter since last week.

That’s what makes real compounding possible. And real scaling.

The flywheel spins on its own. Only infrastructure lets it spin faster year over year - without asking you to push harder.

What got you here won’t get you there

The truth the 80% number forces on all of us:

In the self-serve era, content is what wins.

Not better cold emails. Not slicker GTM engineering. Content - in the broadest sense, discoverability and distribution included - is the ground on which the buying decision actually gets made.

Two things are likely true right now:

1. You’re going to need more of it than you’ve been producing.

2. Every piece has to be dramatically better than what you’ve been publishing.

(I don’t know you, obviously. I’m just speaking to the current overall state of B2B tech content. Most of it is... not great.)

More quantity. Higher quality. Both at once.

That’s not a preference, it is what the 80% demands. The process and infrastructure that got you here won’t get you where you need to be.

A calendar could keep you fed in the old world.

A flywheel could keep you moving through the transition.

But in a world where 80% of the journey is silent and self-directed, you need content treated as a product - researched, created, iterated, launched, distributed, measured, and relaunched - backed by infrastructure, powered by sharp messaging, compounding week over week.

Build the system that will.

‘till later:
Dan Renyi - founder at Klear B2B

Product teams live it - every sprint builds on the last.
Every feature makes the product more valuable.
Every bug fix, every user insight,
every iteration…

They all compound into something that’s worth more this quarter than last quarter.

Finance people worship compounding. The entire venture model is built on it. Every board deck has a hockey stick that assumes it.

But marketing???
Marketing resets, guys.
Regularly.

Not every quarter - that would be too neat and predictable.

Subscribe now

But regularly enough that it makes building and compounding way harder than it needs to be. Here’s what I mean:

- A new CMO brings a new playbook.

- A new CRO restructures the team and the pipeline definitions change overnight.

- the new agency starts discovery from scratch - asking you the same questions the last three people asked.

- Your ICP gets rewritten. Your positioning deck gets shelved. The messaging your team spent four months refining: gone.

Because someone new decided “we need to rethink our story.”

(That’s not a metaphor, by the way. That’s a specific company I’m talking about.)

The gap between how your product compounds and how your marketing resets is becoming impossible to ignore. And AI is about to make it very visible who’s figured this out and who hasn’t.

(Check out the case study about the software team that closed this gap.)

The AI opportunity most people are missing

Right now, the dominant conversation about AI in marketing is about efficiency:

• Write content faster

• Personalize outreach at scale

• Automate reporting

• Generate more output with fewer people

That’s real. It matters, it’s great. But it’s not the big unlock.

The big unlock is this: AI makes it possible for your go-to-market to compound.

Before AI, building a marketing system that actually learns and improves required two things most companies under $10-15M simply didn’t have:

1. Serious headcount - you needed people to maintain the knowledge base, update messaging frameworks, codify what’s working, and train new hires on institutional knowledge and systems

2. Extraordinary discipline - the kind that’s nearly impossible when your positioning gets rewritten every time a new leader joins, and your team gets restructured before you even get used to the last restructure

AI changes both sides of that equation. A single marketing leader with the right AI setup can now build a system where:

• content production draws from validated frameworks instead of starting from a blank page.

• content systems and plans you co-create with the AI actively dig into and feed from living and breathing strategy docs + context coming from sales calls, tickets, reviews etc.

• a new team member or freelancer inherits the full strategic context on day one

• learnings from this quarter’s campaigns actually feed into subsequent decisions

• tone of voice + other brand and corporate stuff get enforced in marketing and messaging

• the list goes on

The difference between “using AI tools” and “building an AI-powered GTM system” is the difference between a faster treadmill and an actual flywheel.

So what does compounding GTM actually look like in practice?

A software product has versions. It gets updated. When something breaks, you fix it - you don’t throw the whole thing out. When someone new joins, they inherit the product as it exists today. They don’t rebuild it from their own assumptions.

Also, it’s impossible for a software product to NOT do what it has been programmed to.

Your GTM should work the same way.

Like… the backbone of it, at least.

People use different terms for this. Some call it a “second brain” for marketing. Others talk about “agentic GTM” or “AI-native marketing operations.” We use the term GTM as Product because it captures the most important part: products compound. Projects restart.

From project mode to product mode

Here’s a quick self-check. Does your company have:

1. A documented system any new hire can inherit on day one - not “go ask the founder” or “shadow Sarah for two weeks and piece it together”

2. A single source of truth for positioning and messaging that gets updated - not a strategy deck from 2024 sitting in a folder nobody remembers the name of

3. Clear metrics that tell you what’s working before you kill it or scale it - not a dashboard that shows impressions and “engagement” while sales complains about lead quality

4. A feedback loop that makes next quarter build on this one - not a clean slate every 90 days where the new plan ignores everything the old plan learned

If you’re missing two or more of these, you don’t have a system. You have a cycle. And that cycle is where the budget goes to die - the expensive reset that happens regularly while your product team ships improvements every two weeks.

The mindset shift that changes everything

There’s an old piece of business wisdom: build processes, not one-off tasks. Systematize what you repeat.

AI takes this to a completely different level. Here’s a concrete example: instead of “write a LinkedIn post” (a task), you build a system that knows your positioning, remembers what resonated last month, speaks in your voice, and learns from every piece of engagement data. The next post isn’t starting from scratch - it’s building on everything before it.

But this requires a fundamentally different way of thinking. And it starts with a deceptively simple question:

“If I had five more people on staff right now, what would I actually be doing?”

Two ways to run with this question. Most people immediately go to interpretation number two:

“More output! Five more sets of hands! We could publish five times as much content, run five times as many campaigns, do outreach to more accounts!”

That’s a good place to start… given that you don’t stop there. Hear me out:

Because that’s exactly what happens when companies actually hire five people without building the system first.

You get five people doing five different things, pulling in five different directions, with five different interpretations of your ICP and positioning. You don’t get 5x the pipeline. You get 5x the random acts of marketing.

The same is true for AI workers.

The real answer is interpretation number one: if you suddenly had seven people instead of two, you’d be forced to build a system. You’d need:

• A documented strategy they can all execute against

• Shared definitions of who you’re targeting and why

• A single messaging framework so everyone tells the same story

• Workflows that connect, not seven people doing seven disconnected things

You’d need the operating system. And the thing about AI is - you can build that operating system now, even with a team of two. The infrastructure that used to require headcount can now be built with structure and the right AI setup.

The people who are getting the most out of AI aren’t the ones producing 5x more stuff. That's almost table stakes, from what I see.

They’re the ones who built the system first and then let AI amplify it. The ones who built the brain before they hired the hands.

Where to start

Pick one process you repeat regularly - content creation, prospect research, campaign reporting.

Ask yourself: “What could I do with 5 extra people?”. Alternatively:
”what would this look like if it got better every time instead of starting from scratch?”

Build the system around that. One compounding process beats ten one-off tasks.

We build this for B2B tech companies in three weeks. The Growth Operating System Sprint finds out what your marketing should actually do, then takes what’s in your head and data - your ICP, positioning, messaging, campaign data, competitive intelligence, brand voice, and GTM logic - and turns it into a living system that your team can run, update, and compound on.

If you’ve been through the reset cycle and you’re ready to break it: Book a strategy call

No pitch deck. Just a conversation about whether this fits where you are right now.

Happy compounding:

Dan Renyi
- founder at Klear B2B, author of this newsletter…

…dad of 4, unofficially the worst cook in his hometown, hobby ultramarathon runner (retired) … - could keep bragging but i’ll spare you.

I‘ve spent the past several months building a system I couldn’t easily explain to people. Not an AI tool. Not a marketing automation platform. Something more like... an operating system for go-to-market.

Then, in the space of two weeks, four unconnected people published content about essentially the same idea - each calling it something different.

Ben Van Sprundel calls it a “second brain.” Maya Voje’s latest Substack calls it an “AI GTM Brain.” Jason Gong at GrowthX is using “second brain” too. Others are just calling it “agentic GTM” and really, dozens of other posts/videos are popping up around the same topic.

They’re all describing one thing. And nobody’s agreed on a name for it.

My reaction wasn’t “oh, interesting new trend.” It was more like: “oh, everyone else is arriving at the same place I’ve been building in.” Which is either validating or terrifying, depending on how you look at it.

(I’m going with validating.)

Let me try to put words to what this thing actually is - because right now the conversation is scattered across different labels, and most of it focuses on only one piece of a much bigger shift.

Subscribe now

One thing worth noting upfront: this isn’t a new SaaS tool. The platform at the center of what we’re building is Claude Code - the same AI coding environment that your engineering team probably already uses to build your software product.

The shift is that this same piece of infrastructure becomes the operating system for your marketing, product and sales team too. You’re not buying new software. You’re setting up, training, and running an environment that already exists - just pointed at go-to-market instead of (or in addition to) code.

The best name I’ve found so far, to capture this thing is GTM as Product.

What “GTM as Product” actually means

Breaking down the buzz and it’s six things:

• Centralized memory and rules. Your ICP, positioning, brand voice, messaging frameworks, strategy - codified in one place that every workflow can access. Not in a Google Doc someone opened six months ago. In the system itself, where the work actually happens.

• Context and intelligence. The system knows your company, your buyers, your market. When it produces output, it draws on real context - not generic best practices it pulled from a training set.

• Predefined workflows. Content creation, prospecting, campaign planning, competitive analysis - these run as repeatable processes with consistent quality. Nobody reinvents the wheel every Tuesday morning.

• Guardrails. Rules that prevent drift. A strategy guard that flags when output strays from your positioning. Quality checks that catch inconsistency before it ships. The system enforces discipline even when the humans are moving fast. (Especially when the humans are moving fast.)

• Direct connections to your tech and data. CRM, analytics, content platforms, project management - the system reads from and writes to the tools your team actually uses.

• Everything tracked and version-controlled in GitHub. Every strategy decision, every workflow, every rule change - tracked, reversible, and auditable. If you don’t know what that means as a marketer, don’t worry - it just means your entire GTM system has a complete history of every change ever made, so nothing gets lost and anything can be rolled back.

That’s it. Those six things together are what makes GTM a “product” rather than a collection of activities people and AI are kind-of-doing.

This is not just automated outbound

One important distinction that most of the conversation is missing.

Right now, the talk around “agentic GTM” is coming from the GTM Engineering crowd and to them, it mostly means automated outbound - AI agents that research prospects, write sequences, send emails.

That’s a real and useful application. But it’s one workflow inside a much bigger system.

What we’re talking about is a platform for orchestrating the entire go-to-market function. Research, strategy, content, product, prospecting, campaign execution, measurement - all running through a system that holds your company’s strategic intelligence.

I wrote about this in more detail in a case study about building a Claude-based Growth OS. In that piece you can see what this looks like when it’s actually running: ű

• AI agents that read product updates to generate marketing content,

• everything stored as plain text files the system can access and update,

• automated visuals, even videos

• a unified workspace where the human reviews and approves rather than creating from scratch.

Mind-blowing, if you think about its implications. And it’s already here, it’s happening.

Share Dan Renyi's Electric B2B

What makes it a “product” and not just “vibe marketing”

There’s a spectrum here.

1. On one end, you have what people are calling vibe marketing - working alongside AI to get a lot done fast. You prompt, it drafts, you edit, you ship. That’s productive. Real improvement over doing everything manually.

2. But take it further. Build actual workflows with defined inputs and outputs. Now you’re systemizing the vibe and adding automation, more predictability. This is where most teams stop, by the way. And it’s where things start to fall apart - because disconnected workflows fade. Someone builds a great prospecting flow, a great content flow, a great reporting flow, and six weeks later none of them are being used because nothing ties them together.

3. Keep going. Add quality gates. Add strategic guardrails that enforce your positioning and tone. Add memory so the system compounds over time instead of starting fresh every session.

4. Then put the whole thing under version control - so every change to your strategy, your workflows, your rules is tracked and reversible. Now your GTM has the same engineering discipline as your product.

Now you have a product. Not a tool you use - a system you own.

That’s what we build for client companies with Growth OS. We set up Claude as the central nervous system for a marketing and sales team - the same infrastructure their engineers already use for product development, now configured with strategy, rules, workflows, and guardrails for GTM.

It’s not new software. It’s a way to train and run an environment that already exists. The kind of setup where you feed it a new competitor and it updates your battlecards, your positioning, and your outreach sequences - because the intelligence is connected.

The part nobody’s spending enough time on

You already feel this, right? The nagging sense that it’s really easy to scale mediocre stuff and bad strategy with these tools?

You’re right.

Here’s the problem with 10x output: it’s also 10x wrong when your strategy is off.

If your positioning is slightly misaligned and you’re producing content manually, you publish 2-3 pieces a week. You notice the drift within a few weeks. Course correction means revisiting maybe a dozen pieces. Annoying, but manageable.

Now run that through an agentic system. Twenty social posts, ten deep-dive articles, a full campaign - all shipped in the same timeframe. Two weeks later you realize your positioning needs adjustment. You’ve got dozens of pieces that thousands of people have already consumed, all pointing slightly in the wrong direction. The cleanup is brutal. Talk about brand confusion at scale.

Same problem with mediocre context and instructions producing and distributing mediocre content.

The faster you move, the more certain you need to be about direction. And you need to be honest about what you’re still uncertain about - because the system will amplify your uncertainty just as efficiently as it amplifies your confidence.

When it all comes together

Having said this - when strategy, positioning, and messaging are locked in, when they’re not in a deck but encoded into the system itself, then GTM built as a product is the most powerful growth engine a B2B company can build right now. Full stop.

This is early. The concept doesn’t even have a settled name yet. Multiple smart people are converging on it from completely different directions. That’s usually the phase right before something breaks through.

If you want to see what this looks like when it’s actually built - not theorized about, but running in real companies - take a look at Growth OS. We go heavy on:

First build the foundations, and then set up the agentic infrastructure on top of Claude Code.

Because the speed is only valuable when the direction is right.

Are you building anything like this? Talk to me.
‘Till next week:

Dan Renyi
- founder at Klear, editor of The Electric B2B Newsletter

Share Dan Renyi's Electric B2B

You know how this goes.

1. Developer finishes the feature.

2. Pings the PM on Slack.

3. PM writes a half-remembered summary.

4. Marketer turns that into website copy.

5. Designer asks for screenshots. Someone records a demo.

6. Someone else updates the docs.

Three weeks later,

• your website still describes the old version.

• Your screenshots show a UI that doesn’t exist anymore.

• Your docs explain a workflow that changed two sprints ago.

Nobody’s wrong here. Everyone did their job.

The process is broken.

Now multiply that by coding agents.

Your engineering team is shipping features faster than ever. Multiple releases a week, sometimes. Heck, maybe multiple a day.

And your GTM team is still doing it the manual way - Slack threads, handoffs, “can someone update the site?”

That gap doesn’t shrink. It compounds. Every feature you ship without updating your marketing is another “lie” on your website that does everyone a disservice. Another outdated screenshot. Another doc that confuses more than it helps because it’s become part obsolete.

Subscribe now

This needs a solution. Yesterday.

And it’s why we talk about treating GTM as product.

Not a buzzword. A real architectural shift. If you think about it:

Wouldn’t your GTM benefit from the coding/product mindset?

What I mean:
You build, version, automate, and deploy your marketing, sales and CS (your go-to-market) the same way your tech team “does software”. So your GTM moves at the same speed as your product.

I don’t think this is going to be a choice, it’ll just be how marketing is done and taught. Because it’s the only way it makes sense. I’ll explain in greater detail through an awesome example…

Nimbalyst is an AI coding environment we use for our “building our GTM product” (I use quotes because it still feels strange to use these phrases) built the clearest example I’ve seen of what this looks like in practice. And once you see it, the old way feels insane.

Here’s how their pipeline works for doing product marketing after a release:

𝟭. 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝘀𝗵𝗶𝗽𝘀. 𝗔𝗜 𝘄𝗿𝗶𝘁𝗲𝘀 𝘁𝗵𝗲 𝗳𝗶𝗿𝘀𝘁 𝗱𝗿𝗮𝗳𝘁 𝗼𝗳 𝗲𝘃𝗲𝗿𝘆𝘁𝗵𝗶𝗻𝗴.

A coding agent reads the feature plan, the actual code, and the release notes. Then it drafts:

• Documentation

• Website copy

• Blog post

• Changelog entry

Not from someone’s memory. Not from a Slack summary. From the source code itself. The team reviews, edits, iterates with the AI until it’s right.

𝟮. 𝗔𝗹𝗹 𝗺𝗮𝗿𝗸𝗲𝘁𝗶𝗻𝗴 𝗰𝗼𝗻𝘁𝗲𝗻𝘁 𝗹𝗶𝘃𝗲𝘀 𝗮𝘀 𝗽𝗹𝗮𝗶𝗻 𝘁𝗲𝘅𝘁 𝗳𝗶𝗹𝗲𝘀 - 𝗿𝗶𝗴𝗵𝘁 𝗻𝗲𝘅𝘁 𝘁𝗼 𝘁𝗵𝗲 𝗰𝗼𝗱𝗲.

This is the decision that makes the whole system possible. Stay with me.

Most companies store their website copy in a CMS - think WordPress, Webflow, whatever. A tool made for humans in mind only.

Nimbalyst’s team stores EVERYTHING - website copy, docs, blog posts - as simple text files (.md to be exact) in the.same.place.their.code.lives. Think about that.

That means the AI that just built the feature can also update the marketing without messy automations that break.

The AI that just built the feature can also update the marketing. No messy automations. No broken integrations. Same environment, same files, accessible by both people and AI through Nimbalyst’s platform (or competing AI code editors like Cursor or Windsurf).

𝟯. 𝗢𝗻𝗲 𝗽𝘂𝘀𝗵. 𝗘𝘃𝗲𝗿𝘆𝘁𝗵𝗶𝗻𝗴 𝗴𝗼𝗲𝘀 𝗹𝗶𝘃𝗲.

IRL what this looks like at Nimbalyst: when a feature is shipped, here is what happens:

• Updated website copy » human review » Automatic push to live site.

• New docs » human review » Automatic push to live site.

• Blog post » human review » Automatic push to live site.

One action. Not five people updating five platforms from Slack threads.

Subscribe now

𝟰. 𝗦𝗰𝗿𝗲𝗲𝗻𝘀𝗵𝗼𝘁𝘀 𝗮𝗻𝗱 𝘃𝗶𝗱𝗲𝗼𝘀 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗲 𝘁𝗵𝗲𝗺𝘀𝗲𝗹𝘃𝗲𝘀.

This is the jaw-dropper. It’s not perfect, but pretty damn good.

They use Playwright to automatically create screenvideos of the new feature in action - a tool that can open your app in a browser and use it like a person would. Click buttons, navigate screens, scroll through workflows.

Their text files include instructions like “capture the dashboard in light mode” or “record a 15-second video of the onboarding flow.” Playwright reads these, opens the app, sets up real data, and captures the visuals.

• Light mode and dark mode - automatic

• Videos with a natural-looking mouse cursor - automatic

• Product UI changes? Re-run the pipeline. Every screenshot and video regenerates. No designer. No re-recording. No “these are from three versions ago.”

𝟱. 𝗧𝗵𝗲 𝘄𝗵𝗼𝗹𝗲 𝘁𝗵𝗶𝗻𝗴 𝗿𝘂𝗻𝘀 𝗳𝗿𝗼𝗺 𝗼𝗻𝗲 𝗽𝗹𝗮𝗰𝗲.

Here’s what ties it all together: Claude Code.

1. The software gets built in Claude Code (via the AI coding agent).

2. The docs get written in Claude Code.

3. The marketing content gets drafted in Claude Code.

4. The human reviews happens in Claude Code and edited in above-mentioned AI editor (Nimbalyst/Cursor/Antigravity etc.)

5. And when it’s approved - it pushes to GitHub, and from GitHub straight to the live website.

The workspace. One approval flow. Build the product, build the story, ship both. That’s not just a set of workflows, folks - that’s an operating system… because, as you will have realized: this product update » marketing pipeline is just part of one segment of go-to-market and there are so many more that you can build on the same philosophy/platform… with the parts interconnected by design, from the start.

Subscribe now

𝗪𝗵𝘆 𝘁𝗵𝗶𝘀 𝗰𝗵𝗮𝗻𝗴𝗲𝘀 𝗲𝘃𝗲𝗿𝘆𝘁𝗵𝗶𝗻𝗴

Our current GTM processes were built for a world where teams shipped monthly. Maybe biweekly.

That world is gone.

Coding agents are shipping features faster than marketing teams can write about them, let alone do launches every once in a while.

The old model - handoffs, Slack threads, manual updates - just breaks. Completely.

You can’t solve that by hiring more marketers. You can’t solve it with better project management. You solve it by treating GTM as a product - something you build, version, automate, and deploy. Just like software.

That’s what Nimbalyst built. Their marketing site is always current. Their docs always match the product. Their videos always show the real UI. Not because someone’s working overtime, but because the system does it.

Here is the deeper breakdown of how this system works on Nimbalyst’s blog.

‘Till next week:

Dan Renyi

ps.: If you’re thinking, “Hmm, wonder how I get started with building a system like this,”… it starts with planning - but not the traditional way…

Jacco and the Winning By Design folks cracked something important with their model, Revenue Architecture. I also think there’s a chapter they didn’t write.

Thanks to Jacco and WBD, have a blueprint for what a revenue factory looks like:

…but we still don’t have one for how the people inside it actually build.

He took manufacturing and systems thinking - production lines, scientific principles, data models - and applied them to recurring revenue. Genuine breakthrough. Long overdue.

But I keep bumping into the layer underneath. How GTM teams actually operate day to day. Whether anything they build compounds - or just resets, or remains isolated.

What I mean is the “visible fractures” you’re familiar with:

SALES

• Discovery questions every rep improvises differently because nobody wrote the good ones down

• Lost deal debrief that happened once on a Zoom call, never heard from again

• Marketing not knowing what objections are killing deals, never making content about them

MARKETING

• A rebrand that never made it into the email sequences, the battlecards, or the onboarding flow

• Personas built in Q1, ignored when writing copy in Q3

• A content calendar full of topics with no thread connecting them to pipeline

The factory blueprint exists. The operating system running inside it is still mostly improvised.

I keep looking at how the best product and engineering teams work - not the theory, the daily practice - and wondering what happens if GTM borrowed that approach. Ownership. Iteration. Building things that actually stick.

I’m starting to think about sales, marketing, CS, and PMM as a product. Something you build deliberately, iterate on, and get better at over time. Still early in this thinking.

Subscribe now

What this looks like in practice

So I’ve been building something. It runs in Claude Code - and honestly the best way I can describe it is: imagine your ICP, your messaging, your competitive intel, your sales playbooks all living somewhere that can actually use them, not just store them.

It requires an additional main layers beneath the “factory” level - that is the buyer journey/data model etc. that sit in Revenue Architecture. So, below the “Factory” we have a “Production” layer.

Part of that production layer (a sub-layer?) is an even more concrete, in-the-trenches aspect (that I haven’t named yet) where the day-to-day work happens with people and tools+agents following processes, projects, one-off things, everything.

Subscribe now

The way it’s structured: Skills (the strategic context - ICP, brand voice, objection library) and Tools. And agents, subagents (that can be wrapped into plugins) do the work.

The nice thing is that it’s flexible. You can use skills/tools in one-off chats with the LLM - and your agents/workflows can use them too.

Now - the obvious question is how this is different from building workflows in n8n or Clay. A few things, but the biggest one is that this is actually interactive.

When you’re running a workflow on an account, a prospect, or a piece of content, it doesn’t just fire and forget. It’ll pause and ask for your input. You go back and forth until the output is actually right - and only then does it move on to the next step, carrying your feedback into everything that follows. That kind of mid-workflow human judgment is just really hard to build into deterministic systems.

That said - it absolutely scales. It can run in bulk, and there are people doing serious outbound volume on Claude Code-based workflows right now, partly because it works well and partly because it ends up cheaper than Clay at scale. So it’s not an either/or.

For my use cases - and the kind of companies I work with, where your TAM might be 400 accounts rather than 40,000 - the interactivity is the thing. Because every workflow needs a bit of actual human thinking, not just a template fired at it.

A few concrete examples of what I’ve been running:

Account research for ABM. Before any outreach, the system builds a brief on the target account - their priorities, content signals, what’s happening in their business - and maps it to your positioning. Something a good marketer would do manually; this makes it repeatable without making it robotic.

Account strategy. Beyond research: who to approach, what angle to take, what to offer at each stage. Pulls in CRM data. Lets you iterate on the strategy before anything goes out.

Sales and marketing alignment. What objections are showing up in sales calls? What content should have existed six months ago? This is the gap that kills most GTM teams - the loop between what sales hears and what marketing makes. This closes it.

All of it connects to the actual stack - CRM, enrichment tools like Clay or Fullenrich, outbound platforms. You plug in what you already use.

Still figuring out a lot of this as I go. Honestly, at the rate of change, this is a perpetual state - we’ll never stop catching up with what’s possible.

But the core idea keeps getting clearer: the problem ain’t that GTM teams lack good frameworks. You spend 10 minutes on Linkedin and walk out with more frameworks than you can count. The problem is that the work marketing+sales teams do doesn’t compound and build on itself the way a good software product does.

That’s the chapter I think still needs writing.

Subscribe now

If any of this resonates - or if you’re already building something like this - hit reply. Always more interesting to think through with people who are actually living this.

Till next week:
Dan Renyi - founder at Klear and host of the recently resurrected Electric B2B Podcast

ps.: Last week, we did a launch of Growth OS, a 2-week strategy+implementation service that
1. plans out your GTM and
2. wires the strategy and foundations into Claude.
The response (measured in pipeline and sales) was way bigger than for any new service/offering I’ve ever launched for Klear over the past 6 years… this tells me I need to go down this path, “productized GTM” is probably where the puck is going.

Today’s issue is different from what you usually get here. It’s more personal, and more promotional. If

• your marketing is working and

• your pipeline is humming along nicely

• or you’re not in the position to change direction

then feel free to skip this one. I’ll see you next week with the usual.

But if you’re running a B2B tech company and you’re the one who MUST figure out growth - this one’s reaaally for you.

Last week, I ran a live session walking through the strategy+execution system I’ve been building. The response genuinely surprised me - I’ve gotten more replies and messages from that session than anything I’ve sent in two years.

• 50+ people registered for the live event (on-purposely left out “AI” from the title to de-hype it). Then 113 people watched the recording the day I sent it out.

• I also put up a resource page with the full recording+slides, prompts, Claude Skills you can use right now. Massive interest, great feedback.

That page comes down tonight, so might want to check/download stuff before it does.

Why this matters

I talk to founders every week who tell me some version of the same thing:

“We don’t have a real marketing strategy. We’re just doing stuff.”

They’ve tried things. Hired a marketer - didn’t work out. Brought in an agency - the agency didn’t understand their industry. Started posting content - got likes from other marketers, zero pipeline.

$150K gone in four months. And deep down, they knew halfway through that it wasn’t going to work. They just didn’t know why.

So they pulled back. “Marketing doesn’t work for us.”

Except it does. It just doesn’t work when it’s built on guesswork.

Here’s what I’ve seen after working with a couple dozen B2B tech companies:

The problem is almost never the people or the tactics. It’s that nobody answered three questions before spending a single dollar:

1. Who exactly are we selling to? Not a vague market segment. The specific segment, event the companies; the exact people inside those companies, and the specific problem that keeps them up at night.

2. What do we say that makes them care? Not your product features. The message that names the problem they’re losing sleep over - in their language, not yours.

3. What are the three things we need to do first - and in what order? Not a 47-page strategy deck. Not twelve priorities. Three moves, validated, prioritized. The ones that will actually move the needle.

Every failed hire or freelancer. Every agency that “didn’t get out industry” Every quarter where marketing felt like throwing money into the dark. It all traces back to those three questions being unanswered.

“I know I need help but I don’t trust anyone to do it right.”

I hear that constantly. (Fair concern, honestly.)

That’s why I built this.

The Growth Operating System is a Strategy Sprint; a focused engagement where I work with you directly to answer those three questions - with data and research, not anecdotes.

We dig into your CRM. We analyze your pipeline - where deals stall, why they die, what actually converts. We look at your competitors from your buyer’s perspective, not yours. We validate your positioning against real buyer language.

And by the end, you walk away with three things:

1. Complete clarity on who you’re selling to, what to say to them, and which channels actually matter for your market.

2. A prioritized 90-day plan - not twelve initiatives, not a strategy deck that collects dust. Three things to execute on, validated against your data, ready to run.

3. A living system that your team can use every day. Not a PDF. The strategy, the messaging, the competitive intel, the plan - all installed as a working system in Claude that keeps your marketing aligned without you having to carry it all in your head.

   Sprint Details

That last part matters more than it sounds. Every $1-8M founder and their brother tells me the same thing:

their best months feel like accidents and their worst months feel inevitable. That’s what happens when the strategy lives in the founder’s head and nowhere else. This fixes that.

Here’s what founders who’ve gone through this process say:

Ahmed Essam, founder at VENTAI: “A masterclass in execution. Dan brings a rare mix of strategic thinking and hands-on delivery.”

Jeff Bates at Trustdock: “Narrowing our focus on a tighter set of accounts and creating targeted campaigns was a game changer.”

Kate Varvedo, Head of Marketing at Infinityn: “We now have positioning and messaging that differentiates us and makes promoting our services more successful.”

The sprint is $3,000 at the launch price. Full money-back guarantee.

Money back guarantee for the whole thing. If you go through the entire sprint, with the Claude-embedded system I build - and don’t feel it was worth every dollar, I’ll refund you in full. And you can keep the AI execution system. I’m that confident in what this delivers - because I’m feeling, living the change it makes.

The launch price ends tonight at midnight Pacific (9 AM Saturday for my European peeps). After that, it goes to $6,000 and stays there.

This is built for you if:

• B2B tech company at $1M-$10M revenue

• 0-5 marketers on staff

• You’ve tried marketing before and it didn’t stick

• You’re done guessing and ready to know exactly what to do next

Book a 30-minute call before midnight and lock in $3,000.

Book strategy call

One (1) spot left for March, 4 for April.

All the best:
- Dan Renyi, Klear B2B

ps.: the call is free, the Sprint will save you quarters of failed experiments and opportunity cost.

I did this live deepdive class last week on why B2B marketing programs break. Spoiler: it is either
1. crappy strategy
2. crappy execution
3. both

And showed people (live) the system I’ve built in Claude that fixes that by… getting rid of the concept of “building strategy” altogether.

I’ve got the edited recording plus some resources I’ve created/added since (like Claude skills, scorecards…)

You can get the recording and resources ungated here.

get all the stuff

Here’s a brief summary:

Most B2B marketing strategies have the shelf life of an open avocado.

You spend three weeks in meetings (or on Zoom) with your leadership team.

You map out the ICP.

You argue about positioning.

You talk about “brand awareness.”

Someone finally exports that beautiful PDF, uploads it to the company drive, and everyone feels like you’ve actually accomplished something. You make vows to follow and revisit the strategy. It’s now your GTM Bible.

Then Monday happens.

Sales needs a new pitch deck.

The CEO sees a competitor on a podcast and wants to know why you aren’t doing that.

Your one junior marketer gets overwhelmed and starts posting “Hot New Feature” on LinkedIn because they don’t know what else to do.

Within 30 days, that $10,000 strategy is a digital paperweight.

I know this because I’ve lived it.

Over the last seven years, I’ve led many deep-dive strategy sprints for B2B tech companies.

Some of them were massive hits - think clients making the Financial Times list of 1000 fastest-growing companies.

Others grew nicely too. And to be honest: there were a few flops in my client engagements, too.

And every single time a plan failed, it wasn’t because the ideas were bad. (I mean, I always have great ideas and I’m sure you do too.)

It was because the plan died on the vine during execution.

I’m writing this because I’ve finally found a way to stop that cycle, and it involves a structural shift in how we think about “strategy.”

=====

You could get the detailed walkthrough and the recording of the class + resources here (it is free and ain’t even gated.)

Get the resources

=====

Why Your Current Plan is Already Dead

If you’re a founder or a sales leader at a B2B SaaS company with a tiny marketing team, you’re likely trapped in the “Random Acts of Marketing” death spiral.

It’s not your fault. It’s a resource problem.

When you’re under-resourced and under-financed, “strategy” feels like a luxury you don’t have time for.

But without it, you end up here:

1. The Research Gap: You think you know your customers, but your messaging is generic. You use words like “seamless integration” and “all-in-one solution” because you haven’t actually mined your sales calls for real buyer language.

2. The Alignment Gap: The CEO wants “brand,” the COO wants “attribution,” and the Sales VP wants “more leads.” Marketing tries to do all three and fails at all of them.

3. The Execution Gap: You have a plan, but no one knows who is doing what on Tuesday at 10:00 AM.

This leads to the “Panic Pivot.”

Results don’t show up in 45 days, so you scrap the plan and try something else.

Repeat until you decide “marketing doesn’t work for us” and go back to 100% outbound cold calling.

The Shift: From “Plan” to “Operating System”

On February 4th, Anthropic released plugins and Cowork for Claude.

Most people saw it as another AI feature.

I saw it as the end of the static marketing plan.

For the last month, I’ve been building something I call the Growth Operating System.

The core idea is simple: Your strategy should not be a document. It should be a living context that sits inside the tools your team uses every day.

Imagine if your marketing “brain”- your ICP research, your brand voice, your competitive intelligence, all the guardrails - lived inside Claude. The tool you use to brianstorm, write and edit.

Now, imagine your junior marketer asks Claude to “write a LinkedIn post about our new feature.”

Instead of spitting out generic AI fluff, the system checks your strategy first.

It sees that your roadmap says: “No feature-led posts until Phase 2. Focus on the ‘Cost of Inaction’ for mid-market CEOs.”

It actually pushes back. It acts as a guardrail.

This isn’t just “using AI.” It’s installing a strategy that can’t be ignored.

How to Actually “De-bullshitify” Your Strategy

One of the biggest problems I see is “Vibe-Based Marketing.”

The founder has a feeling. The marketer follows the feeling.

To fix this, we need to use data—but not the boring “Google Analytics” or a report pulled from Hubspot-kind. I’m talking about buyer intelligence.

Then you apply guardrails: I’ve built a specific Skill called the “Debullshitifier.”

everyone and their brother uses LLMs to put massive amounts of data and ask it to spot qualitative and quantitative patterns and insights. guess what? LLMs are notoriously bad at that.

The Debullshitifier Skill (also accessed here) fixes that.

[Link: You can watch the full 42-minute recording and grab the resources here]

Building the Impossible: A Webinar in One Hour

I wanted to test if this “Operating System” approach actually worked for high-stakes execution.

So, I used it to build the webinar I held itself.

Previously, creating a 60-minute presentation, the slide deck, the landing page for the subsequent offer, and the email sequence would have taken me almost a day of solid work.

Here was the new interactive process, built in Claude Cowork:

1. The Brain Dump: I recorded a messy 10-minute voice note about what I wanted to cover.

2. The AI Iteration: Claude (loaded with my Growth OS context) helped me structure the narrative.

3. The Slide Build: I used a specific skill to generate the actual PPTX slides from the narrative.

4. The Landing Page: Claude wrote the HTML/CSS for the page based on our validated positioning.

I deployed the page also through Claude Code (via the terminal). Insane.

Total time from “idea” to “ready to launch”: About 60 minutes.

Check it out and let me know what you think.

I’ll update you on what’s going on with this platform. In the meantime, do tell me (hit reply or comment) if you have cool use cases for Claude that build growth.

— Dan Renyi

And now it sits.
Untouched.
While the team runs on vibes.

You already know the landscape: channels getting more expensive, cold outreach dying, more competitors than ever,

more confusion than ever, about what to do next.

After 26 research and planning sessions with B2B tech companies over the past couple of years, what clicked for me just a few weeks ago:

planning and execution have always been treated as two separate things

- and that gap is where everything falls apart. The big unlock is merging them so they become almost the same activity. That’s what I want to walk you through today.

How we actually fix this with AI

(and how you can too)

Going into more depth on setting up the agentic strategy+execution layer at our next LiVe cLasS- check it out here.

Our planning process involves three parties working together:

1. The client - knows their company, product, and history better than anyone.

2. The client’s customers - whom we interview regularly. Not surveys. Actual conversations by us, a 3rd party, about how they evaluate, decide, and buy.

3. Us (Klear - and increasingly - our AI agents) - handling research, analysis, and pattern recognition at scale.

If you’re doing this for yourself, you still need three inputs: your internal knowledge, real buyer conversations, and AI-assisted research.

CAUTION: Maybe the most important thing in using AI for strategy: it’s never a one-shot prompt, and not even a prompt-chain, but it’s a back-and-forth, interactive process where we work with our agents to produce the final output. We refine the agent’s output after every step and the “thing” proceeds to the next step using our feedback/modifications. So, human-refined/modified context is key here.

The good news: AI is actually pretty good at self-reflection and catching its own hallucinations and flagging the places where it was just guessing (I taught it to be aware, like a buddhist monk):

Output by Module 5 or our “Growth Operating System” AI thingie - the agent flagged its own guesses and assumptions in the Competitor Research phase.

Here is the process…

Step 1: Build your research agents (not documents)

This is the fundamental shift. Instead of producing a one-time research and strategy document on competitors or buyers, you’re building a reusable research agent that can run again and again every quarter.

Here’s how we do competitor research as an example:

1. [Have context]: competitor research is actually step 5 in our process, where the first 4 steps gather input context and make strategic decisions that are needed for competitor research - and strategy in general. ICP, buyer personas, positioning etc.
   

   

   Green tags = commands that run the step, blue tags=skills that give instruction
   

2. Create Claude Skill(s) that instruct the Agent on the step by step process it needs to follow to map and rate competitors. Tell it to scrape websites, upload context etc.

3. Iterate the Skill’s prompt through conversation. Run it, review the output, refine. Ask it to go deeper on pricing, messaging, positioning gaps. Push back when it’s shallow. Refine until breadth and depth match what you’d expect from a good analyst. This usually takes 3-5 rounds of back-and-forth.

4. Save that final version and context as an agent. Tell it to update itself (the Claude Skill file) based on what it has learned throughout the conversation with you. You now have
   a) a solid competitor research document and
   b) a competitor research tool built specifically for your market that you can run monthly/quarterly to scoop up fresh intelligence.

5. Do the same for:

• Market research - trends, shifts in buyer behavior, new entrants

• Positioning analysis - how your messaging holds up against what competitors are saying

• …and most everything you create inside the strategy.

By the end, you don’t have a strategy deck. You have a set of living research agents that produce living, breathing context and knowledge - calibrated to your specific world. All inside Claude.

Step 2: Execution that feeds off of shared context

This is where planning and execution merge. Whatever you produce through the research/strategy process - positioning docs, messaging frameworks, ICP profiles, competitive analysis - it all goes into a shared context that your AI agents can reference during execution.

Claude Code can now run agents simultaneously - this one researches Substacks and compiles data organized in a spreadsheet for further filtering and actions.

In practice, this means:

1. Store your strategy artifacts as structured documents inside skill files in Claude Cowork.

2. Reference them in every execution task. When you’re writing an email sequence, the agent already knows your positioning. When you’re building a landing page, it already knows your buyer’s language.

3. The context is the plan. And the always-updated research. There’s no separate “strategy doc” that someone has to go read and interpret. The strategy lives inside the system that does the work.

Step 3: Set up continuous monitoring

This is the part that kills the plan-react-panic loop for good. Your research agents don’t just run once - they run on a schedule and flag when something changes.

Here’s what that looks like:

1. Run your competitor research agent monthly. When it detects a new positioning play, a pricing change, or a new entrant, it compares that against your current positioning document.

2. Run your buyer research agent quarterly (or after every batch of customer interviews). Feed in new transcripts and let it pressure-test your ICP assumptions against real data.

3. The system makes recommendations - you can train agents to flag “your messaging around X may need updating because competitor Y is now claiming the same thing” or “buyers are increasingly mentioning Z as a priority, which isn’t reflected in your current positioning.”

4. You decide what to act on. The humans stay in the loop for every strategic decision. The system just makes sure nothing goes stale without you knowing.

Your positioning document from six months ago - the system knows if it’s still accurate. Your ICP assumptions? They get pressure-tested continuously. Nothing rots in a Google Drive folder.

Closed the loop between execution data and strategy

The cool thing is that Cloud Cowork and Claude Code can be used to execute a lot of your GTM plays. This means you don’t have to copy-paste old documents between tools to give other tools context, when creating.

What is possible currently with Claude Cowork and Code?

• Run data scraping and cold outreach in message building? For the most part yes, you can set up workflows for that in Claude. (You’ll need MCP connectors, but still doable.) Guess where the context and the data are going to come from?

• Creating content workflows? Yep, we can rely on all the intelligence of the strategy and research

• Is sales creating messaging and slide decks on an ad hoc basis? You can have Claude check for message consistency and tone of voice, narrative alignment, etc. And Claude now works inside of PowerPoint…

The only gotcha is that your team has to be using Claude for creating content/assets, so that the checking mechanisms work. Where appropriate, you can feed back results from the campaigns or events that you do to the system. You can have skills and agents that will update the existing context and knowledge base

This is what makes it an operating system and not a plan. It’s a living, breathing cycle where strategy informs execution and execution data refines strategy - pretty much automatically - with adult supervision (you).

The tools

Again, this isn’t some expensive tech stack. The whole story runs on Claude (and a few tools connected to it) - specifically Claude Cowork for agent orchestration, and Claude Code for some technical automation. Many of the capabilities we use were literally released on February 4th, 2026. This is weeks-old technology.

Want to see the full system in action?

I’m going live on February 26th to walk through all of this in detail. The strategy components, the agent-setup and this thing dubbed The Growth Operating System. How it actually looks in practice, with live examples.

If you want to see what a living strategy+execution system looks like instead of another dead document - that’s where I’ll show you.

See you Thursday:
Dan

Ps.:Who might you want to forward this to?

Spoiler: this issue will, too, have a happy ending because you’ll learn to spot three silent(ish) killers that have kept you from growing.

This is important stuff: if you regularly launch different “marketing things” like campaigns, outreach, events, content et cetera, but

1. nothing seems to get traction

2. so you end up freezing it…

3. only to try again months later

4. and go “meh” again after some weeks and so on…

So, if a large part of your go-to-market looks like that, you’ll be happy to know there are three underlying forces at play that cause this. If you can spot them, you can fix them.

(Told you this is important!)

The doom loops are three vicious circles that trap small and medium B2B tech companies, especially sales-heavy ones. Most founders are stuck in at least two of them. Many are stuck in all three - and don’t realize they’re connected.

Let’s dive in.

The three doom loops

#1: Random Acts of Marketing

This’ll sound familiar. Leadership has FOMO because (smaller?) competitors seem to be doing a good job, the market is shifting, sales are getting harder and referrals are drying up, so they exclaim:

We need to do some real marketing!

and proceed to dive in and launch without adequate foundations. And that leads to problems:

Each round costs €50-150K. Plus months of opportunity cost.

#2: The Misalignment Misery

Alignment in go-to-market happens (or doesn’t) on multiple levels. Three levels of disconnect that each feel small but together produce total dysfunction:

- Success metrics: CEO tracks revenue. Marketing tracks impressions. Nobody mapped how they connect.

- Sales and marketing: Definitions and process; vague or different ICPs, no shared definition of “qualified,” leads falling into a black hole between teams.

- Story and messaging: Your homepage, sales team, and CS team all tell different stories about what you do. The buyer sees all three. And we’re not even talking about tech documentation yet… And you’re supposed to drive CONSENSUS in the buying committee.

Each misalignment will feel minor. But when they stack? Yuck. Confused output, finger-pointing, and a marketing function that gets cut first when budgets tighten.

#3: The Impatience Tax

Every B2B motion worth doing takes 6-12 months to compound. But the average initiative gets 8-12 weeks before doubt kicks in.

- Something shinier appears. Budget gets redirected. Clock resets to zero.

- Four initiatives × 3 months each = a full year with nothing to show.

- Each failed attempt erodes confidence in the next one.

- So every new try gets less budget, less time, and a shorter leash.

You didn’t fail at any single tactic. You failed at staying with one long enough.

---

These monsters feed each other.

These aren’t three separate problems. They’re one system. Here’s the “playbook”:

**Random Acts creates the mess** (that rhymes!)
No diagnostic foundation means you pick tactics on vibes. Random bets produce ambiguous results. You can’t tell what’s working because you never defined what “working” looks like.

**…which feeds the Misalignment Misery.** Ambiguous results mean every team interprets them differently. Marketing sees “progress.” Sales sees “waste.” Leadership sees “confusion.” No shared scoreboard = no shared reality.

**…which triggers the Impatience Tax.** Misaligned teams + ambiguous results = default move is to kill the initiative. Not because it wasn’t working - but because nobody can prove it was. Clock resets.

**Which brings you right back to Random Acts.** New initiative, because the pressure is real. But same missing foundation. No diagnostic, no shared metrics, no aligned definition of success. Just hope, vibes, and a ticking clock runway.

Round and round. Each cycle more expensive and more demoralizing. And each one making the case for “marketing doesn’t work for us” feel more airtight - when the reality is that marketing never got a fair shot.

I’m running a free live session on February 26 where I’ll show you exactly how to break all three loops - and the system I use with clients to make sure doom loops stay out and “upward spirals” take hold:

I didn’t want to put “AI” in the title, but we’re gonna do some heavy Claude-ing.

---

The root cause underneath all three

I’ve been doing this for over a decade. 7 out of 10 times I troubleshoot why a company’s GTM isn’t working, we end up in the same place.

It’s not a people problem. Your marketer probably wasn’t as bad as you concluded. The agency probably wasn’t as clueless as it seemed. The channel probably wasn’t wrong.

It’s a planning problem. But not “planning” the way you’re thinking.

Not a strategy offsite. Not a ‘Q2 marketing plan’. Not another workshop where everyone agrees on post-its and forgets by March.

I’m talking about diagnostic work that makes everything downstream function. Three things that break all three loops simultaneously:

1. A diagnostic foundation built on data, not assumptions.

- Your CRM tells you how deals actually move - not how you think they move

- Pipeline analysis shows where deals stall and why

- Buyer conversations reveal the language customers actually use (never what’s on your website)

- Competitive analysis from the buyer’s POV - how they actually compare you to alternatives

This breaks Random Acts. When you know what’s happening in your pipeline, channel decisions stop being coin flips.

2. Shared metrics and definitions, agreed before the first campaign launches.

- One definition of a qualified lead

- One shared dashboard

- One set of leading+lagging metrics connecting marketing activity → pipeline → revenue

- Not “marketing metrics” and “sales metrics.” Company metrics.

This breaks the Misalignment Misery. Same scoreboard = no more finger-pointing.

3. Leading indicators set before execution starts.

Not just lagging indicators (revenue, MRR, churn - by the time those move, it’s months too late). Leading indicators:

- Engagement from target accounts

- Reply rates from the right personas

- Qualified conversations started

- Pipeline velocity shifts

These tell you at month 2 whether you’re building momentum. So you’re not making the stay-or-pivot decision on gut feel. This breaks the Impatience Tax.

---

Founders, CEOs skip this work not because they’re lazy. But because it feels slow. It feels like “the stuff before the real stuff.”

But here’s what 10+ years of B2B campaigns have taught me - painfully:

the planning IS the real work.

Everything after it is execution on a clear map. With crap plans, execution is just expensive guessing.

Better plan = less waste.

Better plan = less misalignment.

Better plan = the patience to let the right things compound.

If any of this hit close to home, reply and tell me which doom loop is bleeding you the most. I read every one.

See you next week?

Dan Renyi

PS - If you recognized all three loops running in your company simultaneously... you’re not broken. You’re normal. But normal is costing you a lot. Come to the live session on Feb 26 and I’ll show you how to build the system that breaks all three:

Sign up for session

'done is better than perfect'

for 15 years. Then, my own data just proved me wrong…

I looked back at 10+ years of campaigns and realized that the ones that really worked all had the same thing in common: a foundation that felt almost painfully thorough at the time.

And the ones where we “just got started” because time was tight? Those barely moved the needle.

This pattern shows up every time I troubleshoot why someone’s marketing isn’t working, too. Like, 7 times out of 10, we end up in the same place: the strategy is either non-existent, outdated, or just not thorough enough. The lightbulb goes off for everyone in the room -

“well, if the strategy is crap, what did we expect?”

The bloody meme that immediately comes to mind:

I also realized that I’ve posted way too much about the importance of execution, just assuming that strategies/diagnoses behind them are “good enough” anyways… Now I need to self-correct.

So I built a scorecard. 32 items. Everything I check before launching anything - for clients or for myself:

Share

Click the above image, duplicate the GSheet. Score yourself honestly. Give yourself a “yes” for each item you have documented and up to date. Not “kind of in my head” - actually documented and usable by your team.

By the way, once you conclude your company’s marketing plan sucks is inadequate, you might conclude you need a better strategy. The reality is, you don’t. You need something else:

Subscribe now

Here are some items from the scorecard worth a couple o’ words of explanation:

The Test

Part 1: ICP & Buyer Intelligence (10 items)

2. Negative ICP definition - Who you explicitly do NOT serve. This is the one most people skip - and it’s the reason your pipeline is full of deals that never close.

4. Target account list finalized - Skip this if your ACV is below $100K. Otherwise, an actual list. Named companies. Not “companies like this.”

5. Buyer personas documented - Who inside those companies makes or influences the buying decision? Titles, responsibilities, what they care about. The rest of the items tell you what you need documented for buyer personas.

6. Psychographics - How do your buyers think? What are their values, fears, and motivations? What keeps them up at night - not in general, but about the problem you solve?

7. Jobs to be done (functional, emotional, social) - What’s the functional job they hire you/your product for? What’s the emotional payoff? What’s the social signal? (If you can’t answer all three » do more digging!)

8. Buying triggers - What events cause someone to start looking for a solution like yours? Funding round? New hire? Missed target? These determine your timing.

9. Pain points validated - Not your assumptions. Validated with actual buyers. There’s usually a gap between what you THINK their pain is and what they’d actually say.

10. Objections & blockers documented - What almost stops people from buying? What makes them hesitate? If you haven’t mapped this and if it doesn’t reflect in sales/marketing collateral… no bueno.

---

Part 2: Market & Competitive Intelligence (8 items)

11. Market trends affecting buying behavior - What macro trends are changing how your buyers evaluate solutions? How? AI? Budget cuts? Regulatory shifts? If you don’t know, you’re positioning in a vacuum.

15. Competitor positioning documented - How do they describe themselves? What’s their core message? Where do they show up?

16. Competitor key features/capabilities mapped - Feature by feature. Not to copy them - to understand where you genuinely win and where you don’t. This is actually situation-dependent. I would not dwell on this too much in most cases. The next one’s more important:

17. Your differentiation against each competitor - Per competitor, not generic. “We’re better” isn’t differentiation. “We do X that Competitor A can’t because Y” is. Usually a matrix does it.

18. ”Do nothing” alternative documented - This is your biggest actual competitor and most people never acknowledge it. What happens if the buyer does nothing? What does that cost them? If you can’t articulate this, you can’t create urgency.

---

Part 3: Messaging & Positioning (4 items)

20. ”Problem-Solution Tango” per offering - For each thing you sell: what’s the specific problem, and how does your specific solution address it? One-to-one mapping. If you can’t do this cleanly, your messaging will be vague.

21. ”Why now” articulated - Why should the buyer act today rather than next quarter? This isn’t about artificial urgency - it’s about the real cost of waiting. Ties back into no. 18.

---

Part 4: Metrics & Alignment (5 items)

23. Leading indicators defined - The metrics that PREDICT revenue, not just report it. Pipeline velocity, qualified meetings booked, content engagement from target accounts. If you’re only tracking ARR and churn, you’re driving by looking in the rearview mirror.

24. Lagging indicators defined - Revenue, MRR, churn, LTV. These matter - but by the time they move, it’s too late to diagnose what caused it. You need both.

25. Success thresholds aligned - What does “good” look like for each metric? If your marketing team thinks 50 MQLs is a win and your sales team thinks none of them are qualified, you have an alignment problem.

27. Internal team alignment check - Does everyone - sales, marketing, product, CS - agree on who you’re targeting, what you’re saying, and how you’re measuring success? If not, every department is running a different playbook.

---

Part 5: Execution & Planning (5 items)

28. Buyer journey documented - How does someone go from “never heard of you” to “signed contract”? What are the actual stages, not your CRM’s default pipeline? Where do people stall? Where do they drop off?

29. Map of buyer information needs - At each stage of the journey, what information does the buyer need to move forward? This determines your content strategy. Without it, you’re guessing what to write.

31. 90-day roadmap - What are you doing in the next 90 days, specifically? Week by week. With owners and deadlines. Not a vague “Q2 plan.”

32. 12-month strategy - The longer arc. Where are you trying to get by end of year? What needs to be true for the 90-day plan to make sense?

---

Your Total: ___ / 32

So, you’re suspecting you wouldn’t be scoring very well on this? You might want to attend:

Evaluation

Here’s what I’ve seen after doing this with dozens of companies:

0-10: Flying blind. You’re making every marketing decision on gut feel. Your next hire, agency, or campaign is a coin flip. This is where most companies between $2M and $5M land - and it’s the one reason marketing investments keep disappointing.

11-20: Gaps everywhere. You’ve got pieces of a strategy, but it’s incomplete. Some things are documented, most live in the founder’s head. When someone new joins the team, they spend 3 months just figuring out the basics. Every campaign starts from scratch.

21-26: Solid but not complete. You’ve done real work here. The foundation exists. But the missing items are probably the ones causing specific, recurring problems - like sales-marketing misalignment or campaigns that don’t convert.

27-32: Rare. You have a real strategic foundation. The question at this level isn’t whether you have a strategy - it’s whether it stays alive or decays over time. (More on that soon.)

Most founders I talk to score between 8 and 15. And then wonder why their €150K marketing hire didn’t produce results.

The hire wasn’t the problem. The foundation was.

One more thing.

I’ve been building something that directly addresses what happens after you identify these gaps. Not another strategy deck - something that lives inside your daily workflow and actually stays current.

On February 26, I’m hosting a live session where I’ll show you what this is and how it works. Save the date. Register with a klikk.

Love,

Dan Renyi
- founder at Klear B2B and host of the Electric B2B Show

PS - Reply with your score. I’m genuinely curious where people land.

Even though people have seen the numbers - the undeniable proof that major influence on buyers’ decisions starts way earlier than sales ever gets a chance to pitch - it still helps to provide other angles.

Subscribe now

The age-old question in business is:

"…who do we target, where do we focus our GTM efforts?"

The instinctual answer - even today - leans towards sales-based solutions for most leaders. More SDRs to get the message out, more outreach, more leadgen, more-more-more…

Let’s look at a proven model we have, for a different point-of-view on this. Here is the "demand pyramid" (by the late and great Chet Holmes):

Subscribe now

It's a useful starting point for an overview of buyer readiness in any market. It illustrates the dilemma of "who do we target for immediate pipeline?"

You’ll quickly see that the vast majority of your market is not going to buy this quarter, but they will be in market later.

Showing this to non-marketing executives, someone will inevitably ask:

"What can we do about those people NOW, Watson!?"

…which is the right thing to ask - even better if the question comes from the top.

But... let's not get ahead of ourselves. Here is how the pyramid is built up:

• There are the "very in-market people" who are at the very top of the pyramid - and they're very close to making up their minds.

• Below that is the sweetspot for what you're usually trying to achieve in GTM - buyers who are
  - already actively looking (in market) OR
  - are experiencing pain, (buying triggers activating) but haven't yet started actively looking for a solution.

This is where we want a good part of our focus to go, because this group makes up at least 60% of the market and they WILL buy from our category sooner or later.

It’s all about the message

Also, it's not like you can perfectly target these stages (don't get me started with 3rd party intent signals) so a bulk of what will do the targeting is our messaging.

This is a subtle but big one. As targeting is getting harder because

• of privacy regulations and

• the tools of a multi-billion dollar IT sec industry designed to keep your message off of prospects’ screens, and

• buyers’ ability to smell advertising from further and further away…

…all you’re soon left with is the ability to make an impression with your message only.

That's why it's mega-important to get in the minds of your buyer and whisper the right words. Consistently.

So, what do we do with the bottom 90% of the pyramid?

Back to the dilemma, let’s start with some bad news. In B2B, you can’t make people buy faster. When you’re selling t-shirts and shoes, you can get people to make impulse purchases, but that is not something you can count on in the land of multi six-figure projects and products.

This means that massaging people into a sales-like conversations - no matter how you frame it (“it’s just a chat!” Yeah, right) - is a losing tactic.

The only winning strategy is to build know-like and trust; and stay top of mind. When the time is right, you will be on the vendor short list, ensuring your chance to bid for 77% of the money being spent in your category.

Unless you’re on the shortlist at the beginning of the buyer journey, you’re left competing for the remaining 23% of deals.

Share

Dan Renyi
- Electric B2B