Champion Matrix: Survival to Thrival
Foldspace GTM, internal. Framework: Bob Tinker and Tae Hea Nahm, "Survival to Thrival" (Storm Ventures). Google Doc (v6): https://docs.google.com/document/d/18NjDwdO6fjIGWzBvcqR3JFTRj5Lz226WrfmipVJfjUc/edit — v7, 2026-09-03. Branded PDF: working-docs/champion-matrix-survival-to-thrival.pdf.
Column definitions
- Champion. Who carries us, what they are measured on, whose budget pays.
- Leads (urgent pain). The pain is urgent enough that the meeting happens tomorrow, not next quarter. If they can comfortably put you three weeks out, it is not the pain. For the CTO, add the fear: the fear kills the deal the pain opened.
- Land / Win (date wow). A credible demo that fits their use case exactly and ties straight back to the urgent pain above. Not a product tour: their scenario, their workflow, the pain they just described. That is the TOFU conversion, not a signature. It turns the person into a champion, provided they also believe the company behind it will still be here.
- Marry wow. What gets the ELT to approve: a business case that holds, a vendor that reads as credible and strategic rather than a point tool, and a roadmap they want to be on.
- Expand (1V). The path to standardizing on one vendor for this layer. (1V read as "one vendor" — confirm.)
- 2nd app. The second module attached after the first.
- Hero story. The public story that generates the next wave of leads.
The matrix
| Champion | Leads (urgent pain) | Land / Win (date wow) | Marry wow | Expand (1V) | 2nd app | Hero story |
|---|---|---|---|---|---|---|
| Growth PM — champion, not buyer. Measured on activation, retention, expansion. Budget: marketing, funded by displacing the guided-flow and analytics tooling already in the stack (Pendo, Appcues, Userpilot, Amplitude). Category: product adoption and experience tools, not an AI platform. | Activation and trial to paid flat two quarters, board asking why. Guided flows barely complete, so users never reach the value retention and expansion depend on. Adoption tool renewal coming and its AI module moved nothing. | Their own product, live, in a day, with no eng time and no code changes on their side. User asks for the outcome, product does it, multi screen path disappears. Legible in the category they already buy. | The growth loop is running. The agent surfaces where users get stuck and what they asked for and could not get, ranked by frequency, so she ships against real demand instead of guessing. Remove the friction points, the next cohort converts better, and the loop tightens: intent in, experience out, revenue accelerating. The CTO joins here for the first time, so their date wow runs inside this call: build vs buy drawn out loud, the determinism boundary, roadmap relief. | One flow to the whole adoption surface, then into the existing base for retention and expansion moments. One layer replaces the guided-flow tool plus the homegrown assistant. | Conversational Analytics (GA). Adjacent budget line: every prompt a labeled demand signal, including the unanswered ones. The analytics spend comes across too. | "The adoption stack, replaced." Mixmax retires guided onboarding for an agent: conversion 4.4x, week one engagement 46% to 79%, time to paid 5 to 2.6 days. Logo cleared, runs named. |
| PM — champion, usually the expansion champion. Measured on delivering AI in their area, a data driven roadmap, adoption of what they ship. Budget: same product adoption and experience line plus the "add AI" mandate. | Committed to delivering AI this quarter with no eng capacity. The last thing they shipped was not adopted. Data says what and where, never why; roughly half of product teams have no time for discovery (Atlassian 2026). | Their feature agentified in hours. Ship intents, not screens: existing backend, no UI rewrite. Delivers the AI roadmap item without the sprint they do not have. | The analytics change how the roadmap gets built. Engagement with the agent shows what is actually used, not just what shipped. Intent analytics shows what users are trying to do in their own words (action requests, information requests, reported issues, feature requests), so the roadmap runs on real intent instead of inference from clicks. And the cycle gets short: build, measure, learn, deploy, inside the same quarter. The CTO joins here for the first time, so their date wow runs inside this call: build vs buy drawn out loud, the determinism boundary, roadmap relief. | One feature to the area to every surface. One interface layer instead of an AI experiment per squad; the adoption tool's job moves to the agent. | Conversational Analytics as roadmap input. Dividend, not entry: it reads the Product Agent's conversations, so the agent lands first. Sharpens the roadmap, never writes it. | Optibus: three products in a single sprint, one person on their side. Eitan Yanovsky quote is public. |
| CTO / VP Eng — exec sponsor, usually the economic buyer. Measured on delivering the committed roadmap, build vs buy, the AI cost line. Budget: engineering, defended as roadmap capacity not an AI experiment. | Pain: the roadmap is bigger than the team and features slip, because each one needs a front end first. Fear: nondeterministic behavior in front of their users. An agent doing the wrong thing in a real account is their name on the incident. Fear kills more deals here than price. | The build vs buy line drawn out loud. They keep the loops, models, domain logic, backend. They buy the onboarding and adoption interface that answers user questions in context and is the no-UI path for features that do not justify a screen. Roadmap ships without a front end. | Cost and reliability he controls, with data to act on. Token and action consumption visible per product and per action, so optimizing is a data decision; routing plus real-time context takes 20-35% off the token line. Quota controls for graceful cost safety instead of a surprise invoice. Model high availability, so a provider incident is not an outage. His team runs the Agent Builder itself for instructions and coding: fewer people shipping more. And the boundary holds (APIs, permissions, deterministic system of record, system prompt never exposed). SOC 2 Type II / ISO 27001 / GDPR. | One governed interface layer across products instead of a wrapper per team. Compliance certified once, one eval surface, inference cost behind one gateway. | Trust Lab (early access, forward looking): the answer to the fear, not the entry. Then Conversational Analytics across products. | Optibus: they had the loop and bought the interface. Three products, one sprint, one engineer on their side. |
The two cross-cutting proofs
Land and Marry each need something no single row carries on its own.
Why they want to be your champion (Land). The champion is putting their name on us internally, so they have to believe the company will still be here. What we can say today: two prior exits from the founder (Insightera to Marketo/Adobe, Gainsight PX to Vista Equity), co-author of Mastering Product-Led Growth with Wes Bush, and named customers in Mixmax and Optibus. Gap: there is no funding line documented anywhere in this repo or the bio. Get the exact fact from Mickey before any deck claims one.
What the ELT approves (Marry). Four things, in this order: a business case that holds because the numbers came out of their own environment, a vendor that reads as credible (SOC 2 Type II, ISO 27001, GDPR, references, the action boundary), a strategic buy rather than a point tool (the interface layer, not a feature), and a roadmap they want to be on (Trust Lab, Agent Builder, the analytics dividend).
The CTO is new to the room (Marry). Champion Marry and CTO Land are the same meeting. Whoever runs it owes the CTO their own date wow on the spot: build vs buy drawn out loud, the determinism boundary, and the roadmap relief. Otherwise the loop the champion loves reads to engineering as an unreviewed dependency.
The CTO build vs buy line
| They build and keep | They buy from us |
|---|---|
| The agentic loops, models, prompts behind their domain logic. Layer 1 is theirs. | The interface layer between the user and the product. Layer 2. |
| Backend, APIs, permissions, business logic, system of record. | The onboarding and adoption agent that gets users to the outcome and answers their questions in context. |
| Their differentiated AI features, the ones customers pay them for. | The no-UI path for features that do not justify a screen, so the roadmap ships without the front end work. |
| Their data and the intelligence they build on it. | The conversation record that becomes intent data, which makes the second module obvious. |
Answering the fear, not just the pain: boundary (acts only through their APIs and permission model), scope (actions declared and versioned, not improvised), record (deterministic backend stays the system of record, so a mistake is a reversible action not corrupted state), cost safety (consumption visible per product and per action, quotas degrade gracefully, routing keeps a provider incident from becoming an outage), evidence (Trust Lab evals, early access).
Reading the rows against each other
- Columns are not independent across rows. A champion-led deal reaches ELT approval with the CTO meeting us for the first time. The CTO row's Land beat runs inside the Growth PM and PM Marry beat: one meeting, two jobs. Build the agenda that way.
- Every marry wow is a loop, not a milestone. Growth PM: intent in, experience out, revenue accelerating. PM: build, measure, learn, deploy. CTO: consumption data in, cost and reliability decisions out. The account becomes ours because the loop keeps running, which is also why the second module sells itself.
- Growth PM and PM share a budget and a category. Both bought as product adoption and experience tooling, out of the guided-flow and analytics lines that already exist. Fastest budget in the account because nobody has to create it. The CTO is a different budget and a different argument.
- Champion is not buyer. Line up VP Product / CPO, marketing, or the CTO before the pilot ends.
- The PM is the second module, not the first check. Conversational Analytics needs the Product Agent deployed.
- Language does not cross rows. Funnel language to a PM reads as wrong audience; architecture language to a Growth PM reads as a project she cannot fund.
- The loop closes on hero story. Mixmax carries Growth PM, Optibus carries PM and CTO. Both logos are cleared (2026-09-03), so both run named. Everything else stays anonymized until cleared.
Gaps to close
- The funding line. Nothing is documented in the repo or the founder bio; get the fact before it goes in a deck.
- No published time to first agent metric, so the CTO date wow is qualitative.
- PM quantified proof is thin.
- Trust Lab is early access: it expands an account, it does not close one.