The GTM mistake I made for years before Berlin fixed it.
Why processes outlast people. And what I built from that lesson.
Save this. The processes-over-people section is the one piece worth your Thursday morning.
Send it to 1 founder who is still building around their top performers and quietly wondering why the number drops when that person leaves.
The 90-second version
The origin: GTMcraft started 10 years ago in Berlin, not in 2024. The instinct to coach, document, and systemise without being asked has always been the same.
The system: processes first, systems second, people third. The Winning by Design principle I learned in 2016 and have been building on since.
The asset: two Claude Projects you can set up today: a Sales Coaching System and a Playbook Builder, both described with full prompts below.
The action: run one coaching session or one playbook draft before Friday.
Read time: 9 min
Skip this if you read Monday Tuesday
The GTMcraft system in 5 lines:
Most GTM organisations from 5M to 25M ARR have a systems problem, not a headcount problem
Processes first, systems second, people third. The order most founders have backwards.
Documentation is the foundation AI amplifies. Undocumented processes return noise, not speed
Phase 1 CREATE: build the coaching and documentation foundation before adding people
The two prompts below are the starting point. No community membership needed to run them
Read here the Tuesday edition ↓
Hi, it is Koen Stam and welcome to GTMcraft: The Future 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 scaling 0 to 10M+ ARR.
100,000+ GTM relevant signals from LinkedIn, Newsletters and Podcasts indexed. 100+ playbooks structured. 3 GTM operator jobs across 3 GTM motions (SMB, MM, ENT). Same recipe. Now opening as a community.
2 numbers before we go in
10% of my current paycheck in 2016. That was the year that built the next 10.
100,000+ GTM signals indexed in 2026. That is what the same instinct looks like productized.
Berlin, 2016: Saturdays nobody asked for
I was 29 when I moved to Berlin for the international sales champion role. The title was generous. The actual job was simpler. Take what I was learning from Jacco van der Kooij through the Winning by Design Customer Centric Communication framework and the train-the-trainer program, and coach 6 country leaders and 60+ GTM colleagues in Amsterdam, Berlin, Madrid, Milan, and Paris on how to apply it.
The local sales leaders had the teams. They had the authority. They had the regions. I had a methodology I had just learned myself and a question I kept asking: how do I help these leaders lead better.
Monday to Friday was the assignment. Saturdays were different.
On Saturdays I built enablement content. Outbound playbooks for the BDR teams. Discovery checklists for the new AEs. Coaching frameworks for the country leaders themselves. Onboarding decks for new hires the local teams were about to bring in. None of it was on my objectives. None of it was reviewed by anyone above me. None of it was tracked.
I built it because the work needed it and because building it was the thing I wanted to do that day.
I made 10% of what I make today and I had no clear next step on the org chart.
The instinct that took 10 years to name
At the time I did not have a frame for what I was doing. I thought I was just being thorough. Or maybe ambitious in a quiet way.
10 years later the frame is obvious.
The instinct was to take the messy, lived experience of the country teams and turn it into systems other people could use. Playbook the thing. Document the thing. Build the checklist so the next rep does not have to figure it out from scratch. Coach the leader so the leader can coach the rep. Compound the work.
That instinct is not a career skill. It is the operator instinct. And it does not require authority to be useful. It requires the willingness to do work nobody asked for, in a format other people can use, for a payoff you cannot see yet.
The 6 country leaders did not promote me. They could not. I did not report to them. What they did was let me coach them, trust the methodology, and run the regional motion together. The result was visible. The region moved. ICs got better. Managers got sharper. Leadership saw it.
2 years later that role led to country manager for the Netherlands. 8 years after that, the director international seat I arrived in yesterday.
Nothing about it was linear. All of it was the same instinct.
The thing Winning by Design taught me that most founders still resist
In 2016, Jacco van der Kooij gave me a framework I have never stopped using. The core of it was not a sales methodology. It was a way of thinking about how growth works.
Most GTM organisations are built around people performance. You hire a superstar. The superstar drives the number. You hire more superstars. When the superstar leaves, the number drops and you hire again. The culture becomes the person, not the system.
Winning by Design’s Growth Architecture describes this precisely. SaaS-native organisations stack their model processes first, systems second, people third. The data is based on systems performance, not individual performance. The growth engine runs on the system. People tune it.
I saw this play out in real time across 6 country teams. The markets that were winning had reps executing a common process. The markets that were struggling had a superstar carrying the number alone. When the superstar moved on, the number went with them.
The Saturday work in 2016 was my version of building the system layer. Career frameworks. Onboarding programs by role. SPICED as the shared language from SDR to CSM. Peer coaching structures that let a rep in Amsterdam learn from a close in Milan. None of it was about individual talent. All of it was about making the talent that was already there compound.
That principle has not changed. What has changed is that in 2026, AI makes it non-negotiable.
An undocumented process fed into an AI system returns noise. A documented, structured process returns speed. The operators who have their systems in place are already seeing 30 to 40% output gains from the same headcount. The ones who are waiting to document first are watching those gains go to their competition.
There is no excuse left. Voice notes, structured prompts, AI drafting. The work of documentation is cheaper than it has ever been. The only thing that still requires the operator is the judgment to know what is worth documenting.
I co-lead the Winning by Design Growth Institute chapter for the Netherlands. It is the same conversation every time: founders who know they need to systematise, but have been building around their top performers for so long that the idea of process-first feels like a demotion. It is the opposite. Processes are what make your best people scalable.
What GTMcraft actually is in 2026
GTMcraft is the productized version of the 2016 instinct, rebuilt for the 2026 reality.
3 layers, all working toward 1 outcome: more ARR from what you already have.
Layer 1: Living dynamic GTM Playbooks.
7 GTM playbooks live now. Account prioritization, warm outbound, deal progress and closing, async selling, intent outbound, AI-SDR operating system, international scaling. Each one is a 28-asset suite: the markdown playbook, the master PDF, the 13-sheet workbook, the operating doc, the visuals, the phase guides. The same kind of documentation I was building on Saturdays in Berlin, now built to a tested standard and shipped as a complete operating system.
Layer 2: Group and 1-1 Coaching.
1-on-1 sessions with CEOs and GTM operators. Group coaching cohorts. Async selling cohort running now. The coaching is where the playbooks become decisions. A playbook on the shelf is a document. A playbook in a coaching session is a change.
Layer 3: Claude Projects, Cowork and agents.
Two Claude Projects sit at the core: the Sales Coaching System and the Playbook Builder. Both available to run today. The prompts are below.
All 3 layers point at the same outcome.
The 2026 bet: more ARR from what you already have
This is the part that took me 10 years to be willing to say out loud.
Most GTM organisations scaling from 5M to 25M ARR do not have a headcount problem. They have a systems problem.
AEs who could close 20% more if the coaching was sharper.
SDRs who could book 30% more if the outbound playbook was tighter.
Customer teams who could retain 5 points higher if the renewal motion was documented.
Managers who could move the region if they had a coaching framework that actually got used.
The instinct in 2016 Berlin was: build the system, coach the people, move the region without adding headcount. The bet in 2026 is the same.
More ARR from what you already have. Better systems. Sharper coaching. AI agents that scan the noise so the operator can focus on the decisions.
The founders and revenue leaders I work with are not asking how to hire 10 more reps. They are asking how to make the 8 they have ship more revenue.
GTMcraft is the platform I am building for that question.
Your Claude Projects for coaching and playbook building
Two prompts. Both runnable today inside Claude Projects. Both are the same work I was doing manually in Berlin in 2016. Now with an AI system holding the structure so the operator can focus on the judgment layer.
Prompt 1: Sales Coaching System
Surface: Claude Project (or connect to Gong, Leexi, or any call recording tool to run it as an agent)
What it does: Analyses call transcripts and delivers framework-grounded coaching feedback per coachee, tracking development over time. Runs SPICED or MEDDICCC depending on the organisation. The longer you use it, the sharper the output. It compounds per rep, same way coaching compounds in a real team.
Paste this as your Project instructions:
You are a structured B2B SaaS sales coaching system operated by a GTM coach serving multiple organisations.
Your job: analyse call transcripts and deliver actionable, framework-grounded coaching feedback that compounds over time per coachee.
When a transcript is uploaded, follow this intake sequence:
1. Ask for the coachee name. If known, auto-load their profile. If new, register: name, role (SDR/AE/AM/CSM/Manager), organisation, experience level (junior <1yr / mid 1-3yr / senior 3+yr), ACV band, coaching tone preference (direct / supportive / growth-oriented), sales motion (PLG-assist / SMB velocity / mid-market / enterprise).
2. Ask for call type: Cold Call / Qualification / Discovery / Demo / Follow-Up / Negotiation / Pricing / Closing.
3. Confirm the framework: SPICED or MEDDICCC.
Analysis output structure:
- Executive summary (3-4 sentences, overall impression)
- Framework element scores (1-5 per relevant element, rubric-based)
- Talk-to-listen ratio estimate + question quality assessment
- Top 3 strengths with transcript evidence
- Top 3 improvement areas with transcript moments and rewrite suggestions showing how to say it better
- Next-step quality assessment (was it earned, time-bound, owner assigned?)
- 3 actionable takeaways for the next call
- Reflection questions for the coachee
Coaching tone rules:
- Junior reps (<1yr): supportive, celebrate small wins, frame improvements gently
- Mid-level (1-3yr): growth-oriented, push on patterns, expect self-awareness
- Senior (3+yr): direct, strategic, challenge assumptions, focus on elite-level refinements
After every session, update the coachee development areas. After 3-5 sessions for the same coachee, generate a Development Trend Summary: what improved, what is stuck, recommended focus shift.
Never give generic feedback. Every observation must reference a specific transcript moment. Always provide rewrite suggestions showing how to say it better. Goal: the coachee walks away with 3 things they can do differently on their very next call.What you get: transcript in, structured coaching report out. Every session builds on the last. After 5 sessions for the same rep, you see pattern trends.
What you still own: the judgment layer. Which patterns matter for this rep at this stage. When to push. When to build confidence first. The system holds the structure. You bring the coaching instinct.
One session looks like this in practice. You upload a 40-minute discovery call transcript. The system scores each SPICED element, flags that the rep accepted a surface-level answer on impact without layering a follow-up question, and gives you 3 rewrite options for the exact moment it broke down. The rep walks out of the debrief with 1 thing to try on the next call, not a list of 10.
Prompt 2: GTM Playbook Builder
Surface: Claude Project (or run as a scheduled Cowork session to check for new inputs and build automatically. Details inside the GTMcraft private community)
What it does: Takes your raw operator experience, field observations, and named data points, and structures them into an execution-ready GTM playbook. The same kind of documentation I was building on Saturday mornings in Berlin. Now it runs in 30 minutes instead of a Saturday.
Paste this as your Project instructions:
You are an operator-grade GTM playbook builder for B2B SaaS founders and revenue leaders scaling to $10M ARR.
Your job: take raw operator experience, call notes, and field observations and produce execution-ready playbooks that can be implemented in weeks, not quarters.
Every playbook follows a 3-phase structure:
- Phase 1 CREATE (foundation, Days 1-5): what gets built before anything runs
- Phase 2 OPERATIONALISE (daily execution, Days 6-14): what runs every day and week
- Phase 3 SCALE (compound, Week 3+): what extends the system across the org
When a playbook request is received, ask for:
1. Topic and operator problem (what execution problem this solves)
2. Target company stage (0-2M / 2-5M / 5-10M ARR)
3. Known data points or named examples from your own experience
4. What is working now versus what is breaking down
Output structure per playbook:
- TL;DR (4 bullets: problem / system / asset / action)
- Lessons from the field (first-person, lived operator experience)
- 12-step operating model across 3 phases (4 steps per phase)
- Per step: why it matters, how to execute it this week, what good looks like (4+ observable bullets), common trap with fix
- AI layer per step: what the tool does, where human judgment stays required
- Scaling guidance per step: founder-led / lean team / full GTM
Writing rules:
- Declarative statements only. Cause to effect framing. Specific timelines in days.
- Observable behaviors, not opinions.
- Honest gaps acknowledged when data is not fully there.
- Never use: em dashes, hype words (transform, unlock, supercharge, game-changer), contrast constructions (it is not X, it is Y).
Goal: the operator can implement Phase 1 by the end of this week, Phase 2 by the end of next week.What you get: raw observations in, structured playbook out. The system holds the 12-step architecture. You bring the operator knowledge that makes it real.
What you still own: the anecdotes. The named moments. The specific deals and reps and outcomes that make a playbook credible versus generic. The AI holds the structure. You bring the truth.
Want to run this as an automated agent? The Cowork version checks for new inputs, triggers a build, and notifies you when the draft is ready. That workflow lives in the GTMcraft private community room.
Why I am building this on the side
The honest answer is that I did not start GTMcraft. The Saturday playbooks in 2016 were GTMcraft. I just did not have a name for it yet.
The director role at Personio is the day job. It is the seat where I learn the scale layer of the instinct, work with a senior team I deeply respect, and ship the work that pays the mortgage and funds the side build.
GTMcraft is the creator practice. It is the place where the same instinct that built Saturday playbooks now builds a productized system for other operators. It runs outside Personio. It is not a Personio deployment, not endorsed by Personio, not part of any Personio tech stack.
Both are real. Both compound the same instinct.
If you have ever felt the pull to build something on the side and wondered whether the side work counts: the work you do on Saturdays for the payoff you cannot yet see is the work that becomes your career in 10 years. Document it. Systemise it. Ship it to people who can use it. The instinct is the asset.
What is inside GTMcraft
If this story landed, the place I am building all of it is gtmcraft.circle.so.
7 GTM playbooks live now. The Monday/Wednesday/Friday operator rhythm. All GTM AI agents including the Sales Coaching System and the Playbook Builder above, plus the founder-side Head of Sales on Demand stack. Automated Cowork workflows for both agents. Monthly office hours. Peer matching with operators scaling 0 to 10M+ ARR.
The operator room is open. Operators in the current cohort are using the Playbook Builder weekly and shipping one documented playbook per month.
[DM me to learn more]
2 spots for 1-on-1 work
If the systems problem in your GTM feels urgent and you want to work on it directly, I have 2 spots open for 1-on-1 engagements.
Here is what that looks like in practice. Over the past year I have been working with a sub-5M ARR company. Started as coaching the CEO and one rep on process and methodology. Grew into becoming their Head of Sales on Demand: building the playbooks, deploying the AI agents, and running the strategic GTM decisions alongside their team.
That is the model. Coaching plus agents plus strategic advice, scoped to where your GTM is breaking right now.
If you are a founder or revenue leader under 10M ARR and the documentation and systems layer is the thing holding your team back, reply to this email with one line on where you are stuck. I read every reply.
One question before next week
The Saturday work in 2016 looked unproductive on paper. It was the most productive thing I have ever done.
Reply and tell me: what is your Saturday work right now? The thing nobody asked for, that you are doing anyway, that you suspect compounds.
I read every reply. The patterns in your replies become the next layer of GTMcraft.
See you Saturday in the GTM Signal,
Koen
Your (human) GTM Agent
