The six sessions 🎯 · the spine a vertical roadmap imports
The running situation is Ridgeline Fitness, an invented 22-site gym chain that makes its money on recurring memberships. Twelve candidates go through two gates: whether the business holds the data, and whether the case moves the money that particular business makes. Six clear the first gate. One of the six is worth nothing.
The five categories
See it, explain it, predict it, decide it, do it. A ladder rather than a menu, because each rung eats the one below it. Skip a rung and nothing errors: it ships and becomes confidently wrong.
⚖️ Session 2 · advancedThe four-way verdict
Automate it, augment the person, build something new, keep it human. A separate axis from the category, passed on a task and never on a job, and there is a published argument that the whole field leans toward one of the four for reasons unrelated to which is right.
🎛️ Session 3 · bench nightThe verdict bench
Twelve candidates, five ways to pick, every number computed in front of you. Dropping what cannot be built takes 64 weeks to 27 with the value unchanged, and the most-requested item on the list scores zero.
🚪 Session 4 · advancedThe data gate
Held in full, held partly, not held at all, and the middle one is what wastes projects because the query runs and the answer looks like an answer. Plus how to tell a form change from a standing job before agreeing to either.
💰 Session 5 · advancedThe revenue model decides
The same use case is worth different amounts at different businesses. Also: the three failure-rate statistics everybody quotes to justify a roadmap, and why none of them survives being looked up.
📄 Session 6 · advancedThe roadmap on one page
The artefact a vertical roadmap imports: the money sentence, the sorted list, what is gated and how, the first three things, and what is deliberately not on the roadmap and why. Final scorecard.
Your passport 🎫
Clear all three quiz questions in a session and it gets stamped. Six stamps means you can take a list of twelve wanted things and turn it into a roadmap somebody can defend in a planning round. Stamps live in this browser only.
What this is not, so you land in the right place 🧭
| If your question is | Go here instead |
|---|---|
| How does this map onto one specific industry, end to end? | Ship It: E-commerce Enablers ↗ |
| What does a five-year lookback on a real platform actually find? | Ship It: Five-Year Platform Diagnosis ↗ |
| I have one product, not a portfolio. What do I build and what do I cut? | Product Thinking ↗ |
| Which architecture should this be built on, and what does each choice cost? | System Design & Orchestration ↗ |
| How do I review a data product that already exists? | Data Product Heuristics ↗ |
| How do I tell whether the thing I shipped actually worked? | Experimentation ↗ |
| I have a list of wanted things and no idea which are possible or which pay. | you are in the right place |
This is the shared spine, not a vertical. It deliberately picks no industry, because the whole point is that the four moves are the same and only the answers change. The two vertical courses on the shelf apply it to a specific business model.
Choose your path 🗺️
Three doors into the same conviction: most roadmaps fail on what the business holds and what the business model pays for, long before anybody writes any code.
What is honest here 🧾
Two of the four moves rest on published work that was read at source. One rests on nothing published at all, and session 5 says so in plain words rather than borrowing a citation that does not fit.
| Claim | Status |
|---|---|
| Every figure on the bench | Really computed in your browser from an entity-availability map: a real gate check per case, a real sum, a real unlock pass across all twelve candidates |
| Ridgeline Fitness and its twelve candidates | Invented. Written so the five ways of picking behave differently from each other. Every number describes this portfolio only |
| That the verdict belongs on a task rather than a job | Anchored in a paper read at source which scored 18,156 tasks across 964 occupations and found few if any occupations where every task qualified |
| That the field leans toward automating rather than augmenting | Quoted from the author's own text, read at source, and quoted rather than paraphrased |
| That the revenue model decides the sequence | The playbook's own heuristic, with no published source. A deliberate search found none, and session 5 states that rather than attaching a citation that does not support it |
| The three famous AI-failure-rate statistics | None is stated as fact. One expired by its own terms in 2022, one traces to an unattributed remark on a conference panel, and one is called misguided by a study that went and read its sources |
| Every effort figure on the bench | Stipulated, not estimated. One comparison reverses if one of them changes, and session 3 names which and declines to teach it |
| The five categories | This playbook's own frame. No paper behind it, and session 1 says so rather than dressing it as doctrine |
The knowledge map 🧠
The whole spine at a glance - hover a session to spotlight its concepts, click any node to jump in.