The 90-second version
Q3 closes at the end of this month. Next year’s budget gets drafted off these numbers, and one blended ratio cannot say which channel earned them.
Your target accounts are grouped by industry, vertical or size. None of those is the thing that made anybody buy.
The market says split one marketing leader into 2 hires. On your runway that is the same mistake twice.
The European read: your account count is capped and your senior talent pool is thin. The wrong grouping costs you a year, not a quarter, and you cannot spend your way back to either.
The thread: every decision you make this month is a grouping decision, and the default grouping hides the answer.
Read time: ~10 min
This week’s number 1 move
Split your payback by channel before anyone drafts next year’s budget off one blended figure.
One row per acquisition channel. Founder-sourced, inbound, outbound, partner, events. Per row: fully loaded spend for the last 2 quarters, new ARR from that channel, gross margin, then months to payback. That last one is the channel spend divided by gross-margin ARR, times 12. Sort ascending. Fund the top 2, cap the bottom 1, write the reason beside each.
From my week: on Monday evening I spent an hour with a founder I coach, running at a different scale to mine. Most of his resource pointed at one large customer. The whole call was about isolating that project completely and building a separate motion around it, step by step, rather than letting one account quietly redefine the company.
My take: the number that feels safest is usually the one hiding your best decision. A blended payback tells you the company is fine on average. It has never once told me what to fund on a Monday. Doug Bell put a sharper edge on it this week. Multi-channel carries a cost that never appears as a line item, because every extra channel takes a slice of the same operator. My own rule sits next to it. If growth stops the moment spend stops, you were running a campaign, not building an engine. The only way to tell those apart is per channel.
Do first (90 minutes): export the last 2 quarters from your CRM and the spend from finance. Tag every closed-won deal with its source channel. Build the 5 rows and compute payback. No RevOps? This is 2 exports and a spreadsheet, not a project.
The rule: a metric you can slice beats a metric that is precise. Plan on the channel that pays you back, not the average that hides it.
Sources: Cannonball GTM (Doug Bell), “What Multi-Channel Actually Costs” (September 4); Mostly Metrics (CJ Gustafson), “Why CAC Payback Is More Useful Than LTV to CAC” (August 30); Koen Stam, “If your growth stops the moment your spend stops, you are running a campaign” (LinkedIn, March 2026)
Hi, it is Koen Stam and welcome to GTMcraft OS: The European GTM Operator. This newsletter is built from 100,000+ GTM signals collected from 100+ operators and founders, combined with 13+ years of my own lessons and failures from the trenches. I write at the intersection of go-to-market practice and AI-powered systems for founders and GTM operators scaling their next 2M, 5M or 10M ARR.
Problem 1: your efficiency lives in 1 blended ratio, so you cannot name the lever
From my week: the coaching call ended on one idea. Isolate the motion, then judge it on its own. You cannot do that to a motion you only ever read inside an average.
You report one efficiency number. It looks fine, so next year gets planned on a feeling. The problem is not the number, it is the blend.
CJ Gustafson made the swap this week. LTV to CAC is precise and unusable, because churn, margin and retention all move inside it for reasons you cannot see. CAC payback holds still, and it slices. Doug Bell adds the cost nobody books. Every channel takes a share of the same operator’s week, so a channel that looks cheap on spend can be the most expensive thing you run. Keenan made the same case from the valuation side on Topline. A number produced by heroics is worth less than the same number produced by a system, because only one of them repeats.
My take: I have planned off a blended number and it cost me a quarter of clarity. The average was green the whole way through. Underneath it, 2 channels were carrying the result and 1 had quietly stopped paying me back. My discipline now comes from planning season. Separate what I know from what I hope, then rank every ask by size, cost, impact and likelihood. A blended figure lets you do neither. It has no rows to rank.
The European read: in a second, smaller market a channel does not scale with spend the way it does at home. The addressable account count is a hard ceiling. The per-channel read is the only thing that tells you whether the ceiling or the channel is the problem.
Do first: take the payback sheet from the number 1 move. Circle the channel with the longest payback. That is the line you have been protecting on a feeling.
Do this week: decide in writing whether next year funds it, caps it or kills it. One sentence of reasoning per channel, so the decision survives the meeting where it gets challenged.
Do this month: add operator hours per channel as a second column. Founder read: count your own hours first, because you are the most expensive input you have.
The rule: if growth stops when spend stops, that is a campaign, not an engine. Only the per-channel view tells them apart.
You know it worked when: you can name the 2 channels funding next year and the 1 you capped, without opening a dashboard. Founder read: you stop paying for a channel you kept because it once worked.
The play: 5 steps to find the channel that pays you back fastest (2026 series). Inlined in full below.
Sources: Cannonball GTM (Doug Bell), “What Multi-Channel Actually Costs” (September 4); Mostly Metrics (CJ Gustafson), “Why CAC Payback Is More Useful Than LTV to CAC” (August 30); Topline, “How to make your ARR 20 percent more valuable to investors” with Keenan (August 30)
Problem 2: your target list is grouped by industry, vertical or size, and none of those bought anything
From my week: on Tuesday I sat in a workshop in our office where 35 GTM leaders watched one real account journey. 79 touchpoints, 25 months, 1 closed-won deal. Then the room split into 7 teams to cluster target accounts. All 7 clustered by industry, vertical or company size. Same reflex, same cut, every team.
It is the wrong cut, for a boring reason. Two companies in the same industry at the same headcount buy for different reasons and on different timelines. Two companies in different industries with the same problem buy the same way. The cut that predicts a purchase is the problem being solved.
April Dunford made the adjacent argument this week. When capability moves fast, you anchor on the problem you are trusted to solve, and you stop re-cutting your market every time the category twitches. On 30 Minutes to President’s Club, Armand Farrokh walked his own territory segmentation live. What produced meetings was a repeating situation, never a code on a company record.
My take: the industry cut survives because it is the easiest one to defend in a meeting. Nobody argues with a vertical. The teams I have watched work a list that shared a label and nothing else got their answer from the reply rate long before the pipeline arrived. What I look for now is a repeating trigger: the same event in 3 of the last 5 wins. If I cannot name it, we do not have a cluster. We have a filter, and a filter is not a program.
The European read: an industry cluster here gives you 40 accounts across 6 languages. A use-case cluster gives you 40 accounts and 1 message that travels. That is the only version a team of 3 can run in 4 countries.
Do first: take your last 20 closed-won deals. Write the trigger that started each one, in the customer’s words. Group by trigger and count how many groups cross industries.
Do this week: rebuild 1 cluster around the sharpest trigger. Cap it at 40 to 50 accounts. Sort it into 3 buckets: active focus, future pipeline, and the accounts that define the cluster ICP.
Do this month: put the cluster in front of sales and marketing together, weekly, for 4 weeks. A list nobody revisits is a slide, not a program.
The rule: group accounts by the problem they are trying to solve. Industry, vertical and size describe a company. They do not explain why it buys.
You know it worked when: engagement rises inside 1 cluster before any pipeline shows up, and both teams argue about the same 40 accounts. Founder read: reply rate on that cluster beats your average by enough that you stop guessing.
The play: 5 steps to run a full account program with a team of 3 (2026 series). Takes the cluster from a spreadsheet to a running program on a lean team, with the 3 buckets and the weekly revisit built in.
Sources: Positioning (April Dunford), “Shifting positioning when AI capabilities are rapidly changing” (September 3); 30 Minutes to President’s Club, “My exact prospecting process that got me to President’s Club 3x” (August 25); Outbound Squad, “Tackling new territories, multi-threading and more” with Krysten Conner (August 31)
Problem 3: you are about to buy a role you cannot yet judge, and the market told you to buy 2
From my week: I ran hiring debriefs on a partner-manager role this week. On his second day I handed a new country leader 5 open reqs across 5 functions. Two of those functions I have run myself. The rest I have only ever hired for, which is a different thing, and I try not to forget it.
Kieran Flanagan argued this week that the marketing leader role is splitting into a growth half and a brand half, because almost nobody spikes at both. The trap is reading a market shift as a hiring instruction. Splitting a role does not fix your judgment about the role. It doubles the bet.
Jason Lemkin’s updated list of first sales-team mistakes lands in the same place. His other piece this week draws the line: specialisation earns its place after the motion works, not before. Mark Roberge went further on The Science of Scaling, arguing the shape of the seller role itself is being redrawn. All 3 point the same way. Buy the judgment by the hour first, then hire against a standard you can finally write down.
My take: I have hired for functions I had never run, and the honest version is that I was buying a story I could not test. What I do now is cheaper and much slower to feel good about. I rent the judgment first. Two hours a month from someone who has run that function at my next stage beats 3 extra interviews, because they tell me what to look for instead of who to pick. Headcount is not capacity either. Before I add a seat, I want proof the team I have is aimed at the revenue I already own.
The European read: in a smaller market the senior pool for any one function is a few hundred people who all know each other. You get about one serious attempt per role per year. A bad exit travels faster than a good hire.
Do first: for the next role you plan to open, write the 3 observable things that would tell you someone is good at it. If you cannot write all 3, you cannot interview for it yet.
Do this week: retain 1 advisor who has run that function at your next stage, 2 to 4 hours a month. Hand them the specific decision, not a general remit.
Do this month: build a one-page keep-warm bench. Three to 5 people who are 18 to 24 months from being the hire you want, one line each on what they spike at and what they have never run.
The rule: when you cannot judge the role, do not buy the role. Buy the judgment, then hire against a standard you can write.
You know it worked when: your next senior hire clears a written bar rather than an impression, and the 3 standards are in the scorecard before the first interview. Founder read: your burn multiple holds through the hire instead of stepping up for 2 quarters.
The play: 5 steps to hire the advisor before you hire the VP (2026 series). Puts paid judgment in the seat for a few hours a month, so the standard is written before the salary is committed.
Sources: AI Marketing (Kieran Flanagan), “The CMO role is dying, here is why I think that is temporary” (September 4); SaaStr, “The top 10 mistakes folks make hiring their first sales team” (September 4); SaaStr, “After 2M in ARR, start specializing your sales team” (September 3); The Science of Scaling (Mark Roberge), “The only sales role that will exist by 2030” (September 3)
Save this. 3 GTM problems from my own week, each backed by named expertise and wired to a play you can run today.
Send it to 1 founder or GTM operator who is about to plan next year off one efficiency number.
Also on the radar
A mutual action plan built with someone who cannot commit their own organisation is a to-do list you copied the customer on. A stalling plan then never reads as risk. Action: for your 5 closest renewals, name the 1 person who can commit budget and rebuild the plan with them. The CS Cafe (Hakan Ozturk), “Who you build the MAP with decides everything” (August 30)
A US software tax proposal would add 8 to 10 percent to tools you have already budgeted, in the same planning cycle you are drafting now. Action: list your US-billed line items and re-run next year’s tooling budget with the uplift. SaaStr, “California is taxing SaaS and AI tools” (September 5)
Steal this move: read your payback by channel in 4 weeks
Fixes: the one blended efficiency number that says the company is fine on average and never says which motion to fund.
Best for: 2-5M and 5-10M teams on the new-country or upmarket leg, where 2 or 3 channels carry everything and nobody has separated them.
The 5 steps (over 4 weeks):
Week 1, tag it: pull the last 2 quarters of closed-won and tag every deal with its source channel. Honestly. A deal a founder sourced at an event is not inbound.
Week 1, cost it: pull fully loaded spend per channel for the same 2 quarters, and add the operator hours each one eats.
Week 2, compute it: months to payback per channel. Multiply new ARR by gross margin, divide the channel spend by that figure, then multiply by 12. Sort ascending.
Week 3, decide it: fund the top 2, cap or kill the bottom 1, and write one sentence of reasoning beside each.
Week 4, gate it: set the review date now. A channel gets 2 quarters to move its payback or it loses the budget line.
Template: a 1-5 rubric scored per channel. Rows: source tagging is trustworthy, spend is fully loaded, operator hours counted, payback computed, decision written with a reason. Under 3 on any row means the number in your plan is a guess with a decimal point.
Paste this into your AI:
Here is closed-won by source channel for the last 2 quarters, with fully loaded spend per channel and our gross margin: [paste]. Compute months to CAC payback per channel and sort ascending. Then tell me which 2 channels to fund next year, which 1 to cap, and what my blended number was hiding.
Full play, template and workbook inside GTMcraft OS. Reply or DM me to get access.
Why this matters now
Q3 closes at the end of this month. Next year’s budget gets drafted off whatever these numbers look like, and that deadline is closer than the quarter.
All 3 problems are the same problem in different clothes.
One blended ratio groups your channels so you cannot see which one pays.
An industry label groups your accounts so you cannot see which ones buy.
A market headline groups a job into 2 roles so you cannot see the one you are able to judge.
In each case the default grouping is convenient. Convenient is exactly why it survived.
Cut it once, properly, and next year’s plan starts writing itself.
3 questions for the room
If you could fund only 2 acquisition channels next year and had to cap a third, which 2 survive, and does your payback per channel agree?
What trigger started your last 20 wins, and does your target list group by it?
Which open role can you write 3 observable standards for today, and which one cannot you?
Reply or send me a DM.
Go deeper with GTMcraft OS
Scale on plays and skills, not prompts.
You do not need another newsletter, podcast or LinkedIn post.
You need the system under it.
Read the plays. Every single play is free and live in GTMcraft OS on Notion. Click through from any edition.
Build your own. GTMcraft OS members connect the full libraries, 100+ plays plus 100,000+ LinkedIn, newsletter and podcast signals, straight into their own AI or Claude context, so their team builds on the whole corpus instead of one-off prompts.
The single plays are free to read. Wiring the whole corpus into your own operating system, and keeping it current, is what the membership gives you.
Master your human GTM skills AI can’t replace inside GTMcraft,
Koen
Your (human) GTM Coach
—
PS. Co-written by Wispr + 3 GTMcraft Skills + Claude Opus 5; edited & approved by Koen
The Future GTM Operator is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.