Go-to-Market Before Product-Market Fit: Early Instrumentation That Actually Helps

Founders are told to find product-market fit before scaling go-to-market. That's directionally correct but operationally useless — you can't know you've found PMF without instrumentation that measures it, and you can't build the right instrumentation without understanding what GTM looks like before the fit is obvious. This guide covers the early metrics, survey methods, and tracking choices that help seed-stage teams learn faster — grounded in Sean Ellis's PMF survey methodology and the GTM frameworks popularized by Lenny Rachitsky's newsletter and podcast.

The PMF measurement problem

Product-market fit is not a binary switch. It's a gradient that becomes visible through retention curves, organic growth, and user sentiment — but only if you're measuring the right signals from the start. Most seed-stage GTM instrumentation fails because it tracks activity (signups, page views, demo requests) instead of value delivery (activation, retention, willingness to pay, and the "very disappointed" threshold).

• Vanity metrics — total signups, social followers, press mentions. These feel good and inform nothing.
• Activity metrics — logins, feature clicks, time on site. Useful for debugging UX, not for PMF signal.
• Outcome metrics — activation rate, week-1 retention, NPS/PMF score, revenue per user. These tell you if the product works.

The Sean Ellis PMF survey

Sean Ellis, who coined the term "growth hacking" and ran early growth at Dropbox, LogMeIn, and Eventbrite, developed the simplest PMF signal available: ask users "How would you feel if you could no longer use [product]?" with options ranging from "Not disappointed" to "Very disappointed." The PMF Survey tool at pmfsurvey.com implements this methodology.

The benchmark: when 40% or more of respondents say "Very disappointed," you have strong PMF signal. Below 25%, you have a positioning or product problem. Between 25% and 40%, you're close but need to understand which user segment drives the score and double down on them.

How to run it correctly at seed stage

1. Survey active users only — people who've experienced core value, not signups who never activated.
2. Minimum 40 responses — smaller samples are directional at best; segment by user type if possible.
3. Follow-up question — "What is the primary benefit you receive?" reveals positioning language that resonates.
4. Run quarterly — PMF score shifts as product and market evolve; track the trend, not just the snapshot.
5. Segment results — aggregate scores hide whether enterprise users love you and SMB users don't (or vice versa).

The PMF survey doesn't tell you what to build — it tells you whether what you've built matters to anyone. That's the only question worth answering before you scale GTM spend.

Activation: the metric that predicts everything

Lenny Rachitsky's analysis of activation metrics across hundreds of startups consistently shows that activation — the moment a user first experiences core product value — is the strongest predictor of retention and PMF. Defining and measuring activation correctly is the highest-leverage GTM instrumentation decision at seed.

Defining your activation event

Your activation event is not "signed up" or "completed onboarding." It's the specific action that correlates with retention — the moment the user gets the "aha" experience your product promises. Examples:

• Slack — 2,000 messages sent by a team (historically).
• Dropbox — file placed in synced folder on multiple devices.
• AI writing tool — first document generated and edited (not just opened).
• AI research product — first query answered with cited sources the user saves or shares.

Finding your activation event requires correlating early user actions with 30-day retention. Run a cohort analysis: group users by actions taken in their first session/week, then compare retention curves. The action that separates retained from churned users is your activation event.

The pre-PMF instrumentation stack

You don't need a data warehouse at seed stage. You need five tools configured correctly and a weekly review ritual:

1. Product analytics (PostHog, Mixpanel, or Amplitude) — event tracking for activation, retention, and feature usage. Instrument 10–15 events, not 200.
2. Session recording (PostHog, FullStory, or LogRocket) — watch 5–10 sessions per week to understand where users struggle. Qualitative signal complements quantitative.
3. Customer feedback (PMF survey + in-app NPS or CSAT) — structured sentiment on a regular cadence.
4. Revenue tracking (Stripe dashboard + basic MRR/ARR spreadsheet) — even pre-revenue teams should track willingness-to-pay signals from pilots.
5. CRM or pipeline tracker (Notion, HubSpot free tier, or plain spreadsheet) — every sales conversation logged with outcome and objection notes.

GTM motions before PMF is clear

Lenny Rachitsky's framework for GTM motions identifies four archetypes: sales-led, product-led, community-led, and marketing-led. At seed, you should be running one motion deliberately and measuring its efficiency — not spraying all four.

• Sales-led (founder-led sales) — track conversations → demos → pilots → paid conversions. Measure sales cycle length and objection patterns.
• Product-led — track signup → activation → retention → expansion. Measure time-to-activation and organic referral rate.
• Community-led — track community joins → active participation → product adoption. Measure community-to-customer conversion.
• Marketing-led — track content → traffic → signup → activation. Measure cost per activated user, not cost per click.

Most seed-stage B2B AI products should start sales-led with founder conversations, instrumenting every call's outcome. Product-led growth requires a self-serve activation flow that works without hand-holding — rare at seed for complex AI products.

Weekly metrics review: the ritual that matters

Instrumentation without review is shelfware. We recommend a 30-minute weekly metrics review with founders and product/engineering leads:

1. Activation rate — what % of new users hit the activation event this week? Trend over 4 weeks.
2. Retention — week-1 and week-4 retention for recent cohorts. Is the curve flattening or dropping to zero?
3. PMF score — if survey data is available, current score and segment breakdown.
4. Pipeline — sales conversations, pilot status, and objection themes from the week.
5. One qualitative insight — a session recording, customer quote, or churn interview finding.
6. One decision — based on the data, what will we change this week?

Connecting GTM instrumentation to product decisions

The purpose of pre-PMF GTM instrumentation is to inform product decisions, not to optimize marketing spend. When activation rate is low, the fix is usually product (onboarding, time-to-value, core workflow friction) — not more ads. When PMF score is high among a specific segment but low overall, the fix is positioning and ICP focus — not new features.

• Low activation → product problem — simplify onboarding, reduce steps to aha moment, improve empty states.
• High activation, low retention → value problem — core workflow doesn't deliver sustained value; investigate with churn interviews.
• High retention, low PMF score → positioning problem — users stay but don't consider you essential; sharpen value proposition.
• High PMF score, low pipeline → GTM problem — product works; you need more of the right users to find it.

What to instrument in AI-native products specifically

AI products have unique instrumentation needs beyond standard SaaS metrics:

• Query success rate — what % of AI interactions produce outputs users accept (don't regenerate, edit minimally, or abandon)?
• Time to first value — how long from signup to first successful AI-generated outcome?
• Retrieval quality signal — do users click citations, save results, or share outputs? Proxy for RAG quality.
• Cost per activated user — inference costs during onboarding and activation flow; unsustainable if too high.
• Human escalation rate — how often do users need human support to complete core workflows? High rate = product gap.

When GTM instrumentation is premature

Not every pre-seed team needs a full instrumentation stack. If you have fewer than 20 active users and haven't validated the core workflow manually, skip the analytics platform and talk to users directly. Instrumentation pays off when you have enough volume to see patterns — typically after 50+ users have tried the product and you've manually onboarded the first 10–20.

The sequence we recommend: validate manually with 10–20 users → define activation event from retention data → instrument the 10–15 events that matter → run PMF survey at 40+ active users → weekly review ritual → scale GTM motion that data supports. Skipping steps is how founders build dashboards that nobody reads while missing the signal that PMF is emerging in a segment they haven't noticed.