# PMF Advisor — Signals, Discovery, Pivots, and the Chasm Reference *Part of the PMF Advisor skill: https://wavect.io/.well-known/agent-skills/pmf-advisor/SKILL.md* PMF signals by product category, the customer discovery interview protocol, the pivot framework, and crossing the chasm. ## PMF Signals by Product Category Generic PMF advice is often wrong because the signals differ by category. Always identify the category first, then apply the right benchmarks. ### B2B SaaS Core signal: NRR > 100% and month-6 retention > 40% Supporting signals: - Customers use the product without being prompted by CSMs - Sales cycles are shortening (customers understand the value faster) - Customers can articulate ROI in their own words without coaching - Inbound referrals from existing customers Anti-signals that are NOT PMF: - High activation (people try it) with low retention (they don't come back) - Customers renew because the switching cost is high, not because they love it - Champions leave the company and their replacement cancels - The product requires heavy onboarding by your team to succeed ### Consumer / Mobile Core signal: DAU/MAU > 25%, D30 retention > 20% Supporting signals: - Unprompted social sharing - Negative sentiment when the app is down (people care) - Organic word-of-mouth referral rate > 15% Anti-signals: - High downloads from a PR spike that do not convert to D30 retention - Users open the app but complete the core action only once - Retention driven by notifications / email re-engagement (remove these and watch what happens to retention — this is the real retention number) ### Marketplace Core signal: Liquidity (supply finds demand without manual matching) + repeat purchase rate > 50% on both sides Supporting signals: - Suppliers / sellers join organically (inbound, not outbound) - Buyers return faster than their natural repurchase cycle - CAC for supply is declining as organic supply increases Anti-signals: - You are manually matching supply and demand on every transaction - One side of the marketplace is subsidized (free for buyers or free for sellers) — this masks whether either side has real PMF ### Developer Tools / Infrastructure Core signal: Low churn + community growth + integration depth Supporting signals: - Developers integrate the tool deeper into their stack over time (not just try it) - GitHub stars are growing, but more importantly: issues, PRs, and Stack Overflow questions (indicates real usage, not just interest) - Teams that adopt the tool expand usage to more projects over time - Developers advocate for the tool when changing employers --- ## Customer Discovery Interview Protocol When quantitative data is insufficient, run structured discovery. This is the protocol. Do not shortcut it. **Cadence (Teresa Torres, Continuous Discovery Habits):** Interview at least 1 customer or prospective customer per week, every week. Not in batches. Continuously. The goal is not to "do discovery" — it is to maintain a living model of the customer that is updated weekly. **Interview structure (25 minutes):** 1. **Context (5 min)**: "Walk me through the last time you dealt with [the problem we solve]. Not in general — the most recent specific time." - Never ask hypothetical questions. Ask about the past. - Listen for: timeline, frequency, emotional stakes, who else was involved 2. **Current behavior (7 min)**: "What did you do to handle it? Walk me through each step." / "What tools were involved?" / "What was frustrating about that process?" - Never suggest solutions. Never mention your product. - Listen for: workarounds, emotional frustration, time spent, cost 3. **Stakes (5 min)**: "What happens if this problem doesn't get solved?" / "Has it caused a specific bad outcome? Tell me about that." - Listen for: whether this is a vitamin (nice-to-have) or painkiller (urgent) 4. **Switch signal (5 min, only if they've used a competitor)**: "What made you start looking for a different solution? What was the moment you decided the old way wasn't good enough?" - This is the most valuable question for PMF. The "moment of switch" is where the pain is highest and willingness to pay is highest. 5. **Ideal state (3 min)**: "If this was perfectly solved, what would that look like for you? What would change?" - Listen for: the outcome they want, not the feature they want **Mom Test rules (Rob Fitzpatrick) — apply every session:** - Never ask if they would pay for your product. Ask what they currently pay. - Never ask if they like your idea. Ask about their behavior today. - Never tell them what your product does until after the interview. - Watch for compliments ("that's a great idea!") — they are worthless. Watch for commitments ("I'll pay for beta access") — they signal PMF. --- ## The Pivot Framework The single most misused concept in startups is the pivot. Most founders either pivot too early (out of impatience) or too late (out of sunk cost). **When the data demands a pivot:** - Sean Ellis score below 25% after 3+ iteration cycles focused on the same ICP - Month-3 retention below 20% across all cohorts with no improving trend - No customer can articulate the product's value in their own words after using it - The only retained users are people with a personal relationship to the founders - You have spoken with 30+ ICP-matching prospects and none will commit **When it is fear, not data:** - Sean Ellis score is 30–38% (work the Superhuman method, don't pivot the category) - One cohort has good retention but overall numbers look bad (find that cohort) - A competitor shipped something similar and you panicked - An investor said "the market is small" (investors are often wrong about TAM) **The pivot test:** Before pivoting, answer: "What specific evidence, if I had it, would convince me that the current direction can work?" Then go get that evidence. If it is impossible to get without building more, that is a legitimate reason to consider a pivot. If you just haven't tried hard enough to get the evidence, that is not. --- ## Crossing the Chasm Geoffrey Moore's chasm is the gap between early adopters (who buy because they love new things and can tolerate pain) and early majority (who buy because they trust that proven solutions exist). **Why this matters for PMF:** PMF with early adopters ≠ PMF with the early majority. A product can have excellent retention with early adopters and then completely stall when trying to cross to the early majority. Signs you are hitting the chasm: - Your early retention curves are excellent but new cohorts are declining - The people who love your product are unusually technical or unusually risk-tolerant compared to your stated ICP - Customer testimonials sound like "I can see the potential" rather than "this solved my problem today" - Word-of-mouth is not crossing into mainstream channels **How to cross it:** Target one "bowling pin" segment of the early majority — the most specific sub-segment that has the same problem, the same buying trigger, and where a win creates visible proof for the next sub-segment. Do not try to sell to "the market." Sell to one specific title at one specific company size in one specific vertical. Win there first.