Most pre-PMF founders who reach out for RevOps help are trying to buy their way past a step they have not yet done. The first instinct is a HubSpot instance, a programmatic outbound motion, or a fractional GTM lead. The honest answer is almost always smaller, cheaper, and more uncomfortable, ten conversations with prospects who fit a one-sentence ICP. That is the falsifiable test. Everything else is downstream of it.
Why this matters now
Founders are skipping customer development earlier in the cycle than they used to. AI has compressed the time it takes to ship a website, a demo, a working prototype, and a Series of LinkedIn posts to the point where the gap between idea and "go-to-market motion" feels narrower than it actually is. A First Round Review piece on founder-led growth from April 2025 is blunt about the failure mode, nearly always, a startup's failure has to do with the founder's approach to growth, and the failure starts with delegating that approach before the founder understands their business model themselves. The cost of that delegation is rarely just money. It is six months of optimizing a motion against an ICP that was never validated in the first place.
The trap, hiring outbound or ops before the buyer is validated
Founder-led GTM in the pre-PMF stage is not a sales motion. It is a research motion. The deliverable is not pipeline; it is a defensible answer to four questions, in order, do we have product-market fit; if so, who is the ICP; within that ICP, who is the buying persona; and only then, what is the strategy for reaching them. Skip a step and the spend that follows compounds the wrong assumption.
The most common version of this mistake we see at Checkpoint is a founder with eleven prospect meetings sourced through warm intros, plus a half-built HubSpot, plus an outbound contractor on a three-month engagement. The eleven meetings span gaming, CNC manufacturing, maritime shipping, and a glass-table startup in the US. The HubSpot has properties for a verticalized motion that the call data does not support. The outbound contractor is sequencing into a list nobody scoped against an ICP. The infrastructure is real. The ICP is not.
Why ten meetings is the right falsifiable test
Ten is not a magic number. It is the smallest cohort large enough to surface a pattern and small enough that a founder can run it personally inside six weeks without hiring anybody. The discipline is not the volume; it is the constraint that every meeting is sourced against the same one-sentence ICP candidate, run through the same five questions, and scored on the same three axes. A motion that cannot be falsified at ten meetings will not get more honest at fifty.
The five questions are deliberately boring. Do you have this problem today. What are you doing about it now. What would have to be true for you to switch. Who else inside your company would care about this. Would you take a follow-up call in two weeks. The three scoring axes, does the prospect have the problem; do they pay for a current solution; would they take a follow-up, are the only signal that survives founder enthusiasm.
What a 10-meeting cohort actually proves (or disproves)
Below seven yes-yes-yes outcomes across the cohort, the ICP candidate is not real. The exit condition is not "keep talking to people until the data improves." The exit condition is to redefine the candidate ICP in one sentence and run the cohort again. Above seven yes-yes-yes outcomes, the ICP is provisionally real, and the next investment is sales motion, not tooling, not a CRM rebuild, not a fractional ops hire. The cleanest signal that founders are over-investing too early is the urge to fix the CRM before the prospect list is qualified.
What "yes" actually means on each axis
Problem: the prospect describes the problem in their own words before you describe it in yours. Pay: there is current spend, current internal effort, or a current workaround that has a real cost attached. Follow-up, the prospect will introduce a second person from their organization to the next call. The third axis is the one founders score most generously and the one that matters most. A founder-led motion that cannot generate a second meeting from a first meeting does not have a buying persona; it has a courtesy.
When to bring in implementation help (and when not to)
The honest answer is: not yet. A RevOps engagement at the pre-PMF stage is almost always premature, and we say so. The work we do well, HubSpot architecture, CRM migration, attribution, lifecycle, programmatic outbound, assumes a target buyer specific enough that a workflow can fire against it. Without that specificity, every property, every pipeline stage, and every automation is a guess that will need to be deleted in eight weeks. The cheapest version of help at the pre-PMF stage is an eight-hour ICP and buying persona project, not a three-month implementation.
There is a self-aware version of this advice we owe to founders. Recommending a smaller engagement is not generosity; it is the only honest scope when the underlying ICP has not been validated. We have skin in the game on saying so, the larger projects are commercially better for us, and they are still the wrong call before the ten meetings cohort closes.
Pattern from the field
We recently spoke with the technical co-founder of a DACH-based startup building a developer-tooling data pipeline, two to three months into the company, with a single angel and a presentation lined up at a research-institute partnership track. The team had run roughly a dozen prospect meetings: sourced through warm intros: across half a dozen industries, and was preparing to invest in LinkedIn content and outbound. The honest read on the call data was that no two prospects belonged to the same ICP. The recommendation was not "hire us." It was, pause the marketing investment, redefine the ICP candidate in one sentence, and rerun the ten-meeting cohort against that new candidate. The cheapest unblock at their stage was sequencing, not spend.
Resolution, the ten-meetings playbook
If you are a pre-PMF B2B founder and you are reading this with a half-built HubSpot tab open, run the cohort first.
- Define the ICP candidate in one sentence. Stage, segment, geography, and the specific problem you believe they have. If you cannot fit it in one sentence, you do not have a candidate yet, you have a market.
- Source ten matching prospects. Five from your network, five from one cold channel. Mixing channels controls for warm-intro flattery; sourcing against the one-sentence ICP controls for accidental breadth.
- Run thirty-minute discovery calls with the same five questions. Do you have this problem; what are you doing about it now; what would have to be true to switch; who else inside your company cares; would you take a follow-up. Same script, every meeting.
- Score each meeting on three axes. Problem (do they describe it unprompted), pay (is there current spend or effort attached), follow-up (will they introduce a second person). Yes-yes-yes is the only outcome that counts.
- Set the exit condition before you start. Below seven yes-yes-yes outcomes across ten meetings, redefine the ICP candidate and rerun. Above seven, your next investment is sales motion, not tooling.
- Document the cohort in a single spreadsheet. Prospect, source, ICP fit, the three scores, and a one-line note on the follow-up. The spreadsheet is the artifact that survives the cohort. Treat it as your strategy document for the next quarter.
- Defer the CRM build, the attribution model, and the outbound motion until the cohort closes. If you cannot defer them, you are not running the test. You are decorating the assumption.
Six weeks. Ten meetings. One spreadsheet. The output is either a defensible ICP or a defensible reason to pivot. Either is worth more than the marketing budget you would otherwise be spending against an unvalidated buyer.
Where Checkpoint comes in
If you have closed the ten-meetings cohort with a clean yes-yes-yes signal and the next investment is genuinely sales motion, pipeline architecture, programmatic outbound against a defined ICP, or the operating layer to support the first commercial hires, that is the conversation we are built for. If you have not run the cohort yet, the most useful thing we can do is point you back at it. Talk to us when the cohort closes, not before.
Sources
- Lerner, M. "You’re not ready for a head of growth: run this founder-led growth playbook instead." First Round Review, April 2025. review.firstround.com
- Lemkin, J. "How to transition from founder-led sales: why most get it wrong and how to get it right." SaaStr, August 20, 2025. saastr.com