Bridging the Viability Gap: A PMM's Guide to Putting ICP Research at the Heart of Product Development
TL;DR — The gap between what Product builds and what the market actually wants is not a communication problem. It is an infrastructure problem. ICP research — Jobs-to-be-Done, win/loss analysis, search intent data — needs to be a direct input to the product roadmap, not a post-launch messaging exercise. This is how PMM Leads close the Viability Gap before it costs a sprint cycle.
The Viability Gap: Where Product Investment Goes to Underperform
There is a specific type of product failure that does not show up in engineering post-mortems because the engineering was not the problem.
The feature shipped on time. The code was clean. The QA passed. The launch happened. And then — almost nothing. Usage was low. Sales found it hard to pitch. Customer success fielded confused questions from users who did not understand the use case.
This is the Viability Gap: the distance between what the product team built and what the market was actually trying to accomplish.
It is the most expensive failure mode in B2B product development, and it is almost entirely preventable. Not through better engineering. Through earlier, more rigorous, more structurally embedded ICP research.
The current model has research happening too late and landing in the wrong place. Win/loss interviews become positioning documents. User research becomes UX improvements. Competitor analysis becomes a battle card. All of this is useful — but none of it is feeding the roadmap. None of it is shaping what gets built before the build starts.
That is the gap. Here is how to close it.
Why ICP Research Is Treated as a Messaging Exercise
Before addressing the fix, it is worth understanding why the current model persists.
PMM is structurally positioned downstream of Product in most organisations. The PMM Lead receives a feature brief, builds the messaging, trains Sales, and executes the launch. Research, in this model, is input to the message — not input to the product decision.
This positioning made more sense in an era when product cycles were long, research was expensive to conduct, and the PMM function was narrowly defined as launch execution. In a world of continuous delivery, shorter cycles, and agentic tools that compress the time from research signal to synthesised insight, the argument for keeping PMM downstream of product decisions does not hold.
The PMM Lead has access to signals that the product team does not: what buyers say in discovery calls, what deal blockers appear in win/loss analysis, what search terms the ICP is using to describe the problem they are trying to solve. These are not messaging inputs. They are product inputs. And they need to flow upstream.
The ICP Research Stack for Roadmap Influence
Jobs-to-be-Done as a Prioritisation Framework
Jobs-to-be-Done (JTBD) is the most powerful framing tool for translating ICP research into product language. The core question — what job is the customer hiring this product to do? — cuts through feature speculation and anchors the roadmap in the outcome the buyer actually cares about.
In practice, this means conducting JTBD interviews with the right cohort: recently converted customers, churned accounts, and active prospects who chose a competitor. The questions are not about product features — they are about the moment of decision, the struggle before adoption, and the outcome the customer is measuring.
The output of a rigorous JTBD programme is not a persona document. It is a ranked hierarchy of jobs — the outcomes the ICP is trying to achieve, ordered by frequency, urgency, and the degree to which current solutions (including your product) fall short. That hierarchy is a direct input to roadmap prioritisation.
Win/Loss Analysis Beyond the Anecdote
Most organisations do win/loss analysis. Very few do it in a way that generates actionable product signal.
The common failure mode: win/loss data is collected informally through sales team memory, aggregated into a quarterly slide, and used to update battle cards. The feedback loop to Product is indirect at best — a slide presented in a leadership review, noted, and rarely acted upon.
A structured win/loss programme — consistent interview cadence, recorded and transcribed, synthesised against a standard taxonomy — generates something qualitatively different: a pattern database of the specific capability gaps, friction points, and unmet expectations that are driving competitive loss. When that database is queryable, the PMM Lead can surface precise, evidence-backed prioritisation arguments to the product team.
Clay is particularly useful here for enriching the win/loss dataset — overlaying firmographic and technographic data on deal outcomes to identify whether capability gaps are ICP-specific (a problem for a particular segment) or systemic (a problem across the board). That distinction changes the product recommendation significantly.
Search Intent Data as Demand-Side Evidence
Google Search Console is the most underused ICP research tool available to most marketing teams, primarily because it is treated as an SEO asset rather than a market signal.
The queries that bring prospects to your content are a direct window into how the ICP describes their own problem — in their words, at the moment of active search. When a significant volume of searchers are arriving via a query that describes a problem your product does not yet solve, that is not just a content opportunity. It is a demand-side signal that belongs in a product discovery conversation.
The workflow: export Search Console query data monthly, filter for queries with meaningful volume and relevance, and run a synthesis pass using Claude Code to identify the emerging problem categories the ICP is describing. Flag any categories that represent unmet demand — problems the search volume proves are real, but which your current product does not address. Route those to the product team as prioritised input, backed by search volume as a proxy for market size.
This is Agentic Workflows applied to market intelligence — automated, continuous, and feeding directly into GTM Infrastructure and product decision-making simultaneously.
Building the Upstream PMM Presence
Closing the Viability Gap is not just about collecting better research. It is about repositioning the PMM function structurally within the product development process.
Practically, this means two things:
1. A recurring research-to-roadmap ritual. A monthly or quarterly session where the PMM Lead presents ICP signal — JTBD findings, win/loss patterns, search intent trends — directly to the product team, framed as prioritisation input rather than as messaging feedback.
2. A shared taxonomy. Product and PMM need a common language for describing customer problems. When JTBD findings are expressed in the same terminology that appears on the product backlog, the translation layer disappears and the signal flows cleanly upstream.
Product-Led Growth companies have an inherent advantage here: the product itself generates behavioural data that maps onto JTBD. But even in sales-led motions, the ICP research signals are available. The PMM Lead simply needs to route them upstream rather than converting them into messaging collateral and stopping there.
Key Takeaway for PMM Leads
ICP research is not the raw material for positioning. It is the raw material for product decisions — and positioning. The PMM Lead who builds a structured pipeline from JTBD interviews, win/loss analysis, and search intent data into the product roadmap conversation is not doing extra work. They are doing the work that prevents the Viability Gap from appearing in the first place. Route your research upstream, not just downstream, and you change what gets built — not just how it is described.
Frequently Asked Questions
How do you get product teams to actually act on PMM research?
Framing and credibility. Product teams respond to evidence that is specific, quantified, and expressed in language that connects to the problems they are already trying to solve. “Customers want better reporting” is not a product input. “Seventeen of our last twenty-two churned accounts named the absence of role-based reporting access as a primary reason for switching, and the top search query driving traffic to our pricing page is ‘reporting permissions by user’” is a product input. Specificity and evidence get the seat at the table.
Is there a risk of PMM overreaching into product decisions?
The distinction is input versus decision. The PMM Lead’s role is to surface market signal with rigour and route it to the people making product decisions — not to make the roadmap calls unilaterally. Done correctly, this extends the product team’s market intelligence without encroaching on their authority. The risk of overreach is real, but it is managed through clear positioning: PMM owns the demand signal, Product owns the prioritisation call.
How does Clay help with ICP research beyond win/loss enrichment?
Clay’s data enrichment capabilities — pulling firmographic, technographic, and intent data at the contact and account level — allow you to build a richer picture of your ICP across the entire funnel, not just in win/loss scenarios. Enriching your active pipeline with technographic data, for example, can reveal whether a particular capability gap is consistently appearing in accounts with a specific tech stack — which is both a sales signal and a product signal.
How often should the ICP research feed be updated?
The win/loss cadence should be monthly at minimum — enough to catch emerging patterns without being so frequent that individual data points dominate the signal. Search intent data should be reviewed monthly. JTBD interviews are higher investment and can run quarterly, though the synthesis should be updated whenever a material new pattern emerges from the ongoing conversation.