Can an ICP diagnostic be done in less than a month?

Yes. The diagnostic in this case study completed in five business days due to launch timing constraints. Speed depends on access to stakeholders, availability of product and competitive data, and the complexity of the market. The methodology is structured to extract signal efficiently — what determines the timeline is access, not the work itself.

CASE STUDY | AI DEVELOPER TOOLING COMPANY — CORPORATE SPIN-OUT

From 10 Definitions to One Direction

What Happens When Nobody Agrees on the Customer

March 2026

The Setup

Twelve people. Two product launches in six weeks. Seven years of the parent company's history — and zero revenue to show for it.

The company was a corporate spin-out: a new entity built on technology developed over nearly a decade by an established software company. The product was a developer tool for building and debugging AI pipelines — technically strong, genuinely differentiated, and completely unclear to anyone outside the engineering team.

The team was preparing for two launches: an open-source community release followed by a commercial SaaS product. They had a CEO based in Europe, a COO managing operations in the US, a technical co-founder, and a small team of engineers and developer relations. They had budget from the parent company. They had a launch date. They had a product.

What they did not have was an answer to the most basic question in their go-to-market: who is this for?

Not "what market are we in?" They could answer that. Not "what does the product do?" They could demo that. The question they couldn't answer — and hadn't realized they couldn't answer — was: which specific person, at which specific company, with which specific problem, would use this product this week?

That question had never been committed to. And because it hadn't, everything downstream — messaging, launch strategy, pricing, success metrics, even what to say at the launch event — was provisional.

The Diagnostic

We ran a structured diagnostic across five independent sources. Not to collect opinions about what the ICP should be — but to find where the evidence converged.

An anonymous team survey. Ten team members answered the same set of questions about who the customer is, what the product does, and what success looks like. The results were the first signal that something structural was wrong.

One-on-one stakeholder sessions with leadership, engineering, finance, and developer relations. Structured to extract specific, deliverable-ready data — not open-ended opinions.

An independent product assessment. Not just the pitch deck — the actual codebase. We had a developer in the target audience test the product against what the marketing materials claimed it could do.

Competitive landscape analysis. Mapped every competitor to identify the defensible white space the product could occupy — if the team committed to it.

External validation. Tested the emerging ICP hypothesis against real developers who matched the target profile, to see if the product clicked for the person the evidence said it should click for.

The methodology is simple in principle: five independent sources must converge before a recommendation is committed. This prevents the engagement from simply reflecting back what leadership already believes.

The Patterns

The anonymous survey asked each team member a straightforward question: who is the customer?

Ten people gave ten different answers.

Some said senior engineers. Others said junior developers. Some said data scientists. Others said platform teams at enterprises. One said "anyone building AI applications." Another said "developers who are frustrated with existing tools."

These are not minor variations. A product positioned for senior engineers at enterprises has a fundamentally different narrative, pricing model, and go-to-market motion than one positioned for junior developers learning AI. The team had been executing across all of these simultaneously — which meant none of them were getting the focused effort required to actually convert.

This is not unusual. In every engagement at this stage, the question "who is your customer?" produces as many answers as people you ask. The problem is not that nobody has an opinion. The problem is that no single definition has been fully defended, adopted, and enforced across the organization. Without that, every team member optimizes for a different buyer — and the effort fragments instead of compounding.

Ten People, Ten Definitions

The independent product assessment — testing the actual codebase, not the pitch deck — revealed something the team had missed.

The independent product assessment revealed something the team had missed. The product had capabilities that no competitor in the space could match — broader integration coverage and deeper technical functionality than anything else available. But it also had marketing claims the product couldn't support: features referenced in documentation that had no implementation code behind them, a deployment option that didn't exist despite being described in the setup guide, and a quickstart guide with critical errors that would have failed any first-time user attempting to install the product.

This pattern — genuine technical differentiation buried under unverified claims and broken first impressions — is common in technically complex products built by strong engineering teams. The engineers know what the product can do. The marketing materials describe what they wish it could do. And the gap between the two creates credibility risk at the exact moment of highest visibility: the launch.

The quickstart errors alone would have damaged every first interaction with the product. A developer who downloads a tool, follows the getting-started guide, and hits a broken install command in step three does not file a bug report. They close the tab.

The Product Was Stronger Than the Team Realized

During the stakeholder sessions, a specific pattern emerged around how decisions were being made — or more precisely, how they weren't.

The company had recently changed the product name. The decision was made informally in a conversation, without documentation of who approved it, what the reasoning was, or what downstream work would be affected. That single undocumented decision cascaded into over $15,000 in rework: legal filings, branding assets, domain registration, updated marketing materials, team time spent asking "wait, what's the new name?" across multiple channels.

This was not an isolated incident. It was the pattern. Decisions were made conversationally, reversed when someone new had an opinion, and never formally committed. The result was an organization that was perpetually rebuilding the foundation it had just laid — consuming energy without producing progress.

The $15,000 Decision Nobody Wrote Down

The team was weeks away from a launch event. They had a venue, a date, speakers, a demo, and catering. What they did not have was a measurable definition of success.

When we asked what a successful launch would look like, the answers ranged from "lots of people show up" to "we get GitHub stars" to "people are excited about the product." None of these are measurable. None allow the team to evaluate, after the event, whether the launch actually worked.

Without success criteria, a launch cannot fail — which means it cannot teach you anything. The team would walk away from the event with impressions, not data. And impressions are the raw material of the next relitigated decision: "I thought the launch went great." "Really? I thought it was mediocre." Another debate. Another meeting. Another month of provisional effort.

The Launch Without a Scoreboard

What the Diagnostic Produced

In seven business days, the engagement delivered three things:

A unified ICP with evidence. Five independent analyses converged on the same target: a specific seniority level, experience range, and use case that the team had not been targeting. The recommendation came with the reasoning, the evidence table, and the explicit deprioritization — which audience to stop pursuing and why. The CEO, before meeting with us directly, independently confirmed the findings aligned with his own strategic concerns. The ICP was not a surprise. It was a confirmation that the team had been avoiding the commitment.

Measurable launch success criteria. A single-page framework defining what success looks like across five categories — attendance, product engagement, community growth, content visibility, and feedback quality — with floor, target, and stretch levels for each. Designed to be reused for subsequent launches.

A decision map. Who owns which decisions, who approves, who contributes, who is informed. Built from the materials and dynamics observed during the engagement — not theoretical, but descriptive of how the company actually operates.

The Takeaway

The hardest part is not finding the answer. It is making the commitment when you still have uncertainty.

This engagement delivered a 10-20x return on the fee — measured in prevented waste, identified product risks, and unified strategic direction. The ICP work alone collapsed months of internal debate into a single, evidence-backed commitment. The product assessment caught errors that would have damaged every first interaction with the product at launch.

But the diagnostic also revealed something the deliverables couldn't fix: the organization lacked the capacity to implement what it already knew was right. The ICP was correct and everyone agreed — but the event messaging didn't change. The success metrics were defined — but nobody tracked them. The decision map was built — but decisions continued to be made informally and reversed.

This is the pattern that separates companies that convert traction into growth from companies that stay stuck: it is not the quality of the strategy. It is whether the commitment survives the next week of execution pressure, new data, and conflicting input.

The diagnostic produces the decision. What determines whether the company compounds from it is whether the organization — the people, the processes, the habits — can actually hold the line.

For companies that can: the engagement accelerates everything. A wrong decision tested with full commitment teaches you more in 90 days than three uncommitted directions teach you in a year.

For companies that can't yet: the diagnostic still has value. It names the decision that needs to be made. It shows what the evidence supports. And it makes the cost of not committing visible — so that when the organization is ready to hold the line, the direction is already clear.

Results at a Glance

ICP alignment: 10 conflicting definitions → 1 committed, evidence-backed direction
Delivery timeline: 5 business days (vs. typical 4-6 week engagement)
Product blockers caught: 3 critical quickstart errors identified before launch
Unverified claims flagged: 4 marketing claims with no implementation code behind them
Decision cost identified: $15,000+ in cascading rework from a single undocumented decision
Validation approach: Multiple independent analyses converged on the same recommendation
ROI: 10-20x return on a single-phase engagement

Does This Sound Familiar?

The patterns in this case study — fragmented ICPs, unverified product claims, undocumented decisions, launches without scorecards — are not unique to this company. They are structural to the transition between early traction and compounding growth in technically complex B2B products.

If your team has real traction but growth isn't compounding, the diagnostic process described here is available as a structured engagement. It starts with one question: can your team give the same answer when asked who the customer is?

If the answer is no — or if you're not sure — that's where we start.

Get a Clear Read on Your PMF

Frequently asked questions

Everything downstream fragments. Messaging speaks to multiple audiences and resonates with none. The product roadmap tries to serve competing use cases. Sales conversations start from scratch every time because there's no consistent narrative. Marketing generates interest that doesn't convert because the buyer being attracted isn't the buyer the product is built for. The team works harder every quarter without growth compounding, because effort is being distributed across uncommitted directions instead of concentrated on one.

Ask five people on your team who the customer is. If you get five different answers — or answers phrased as market categories rather than specific job titles at specific company types — that's the signal. Other symptoms include sales cycles that require custom explanations for every prospect, marketing that generates interest but not conversion, internal debates about product priorities that never fully resolve, and a pitch that changes depending on who's in the room.

Because the technology was built inside a parent company's context — for that company's customers, with that company's distribution, against that company's competitive landscape. When the product spins out, the technology transfers but the go-to-market assumptions don't. The spin-out team inherits a product without inheriting the buyer clarity that the parent company had (or thought it had). The result is a working product with no committed answer to who it's for in the new context.

Without success criteria, a launch cannot fail — which means it cannot teach you anything. The team walks away with impressions instead of data, and impressions become the raw material for the next unresolved debate. One person thinks the launch went well, another thinks it was mediocre, and the conversation loops without resolution. Defining measurable criteria before launch turns an event into an experiment that produces actionable signal regardless of the outcome.