Cultivate Method
Sign inRequest access
For product leaders

What Cultivate looks like in a real product team

If your product organisation has more than one PM, more than two squads, and more than a year of accumulated written evidence, Cultivate is built for you. Here's what running it actually looks like across the roles in your team.

If you're a Head of Product / VP Product

  • A single source of truth your strategy work earns its keep on. The five baseline files (product proposition, vision and strategy, lifecycle model, target markets, constraints and assumptions) are the strategic spine; everything downstream is evaluated against them.
  • Cited evidence behind every recommendation— the answer to “why are we shipping this?” is no longer a memory test. Click into the recommendation, see the parent opportunity, see the customer evidence the opportunity was built on. Three clicks; no Slack archaeology.
  • A defensible decision log for board meetings, exec reviews, and post-mortems. Every reconcile-driven baseline change carries a dated changelog entry naming the artifact that triggered it; historical positions are annotated as superseded, not deleted.

This role's reading: see Why opportunity-first and Why baselines are sacred on the methodology page — those are the principles that determine whether the discipline fits how you already think about strategy.

If you're a Product Manager

  • A wiki that updates itself as you drop in research notes, call transcripts, support summaries, and competitor scans. The synthesis pass picks the right pages to update and proposes the edits; you review the diff before anything sticks.
  • Recommendations that come pre-bundled with their evidence and constraints. Paste the export-to-prompt brief into Claude Code, Cursor, or Lovable; get a working prototype the same afternoon, with the strategic guardrails baked in.
  • A clear opportunity → recommendation → delivery thread that survives planning cycles and re-orgs. When the same problem shows up in sales notes, support tickets, and research three quarters running, the wiki notices.

If you're a Product Researcher / Insights lead

  • Your research lands somewhere it gets read instead of archived. Drop a transcript in; the LLM tags it against the existing customer / problem / insight pages and surfaces the cross-team patterns immediately.
  • Cross-team patterns surface automatically. When sales, support, and research all flag the same theme over a quarter, the wiki connects the dots — the team doesn't have to wait for someone to notice.
  • Strategic insights compound rather than getting re-discovered every quarter. The wiki is the institutional memory the org-chart keeps disrupting.

If you're a Product Designer

  • The brief you receive carries the why and the guardrails — cited evidence (the customer pains the work is responding to) and the strategic constraints it has to operate inside — not just the what.
  • AI prototyping tools have the context they need. Your design exploration starts further upstream. Claude Code or Lovable receives the recommendation, the parent opportunity, the evidence, and the foundations — labelled as non-negotiable. The prototype reflects the constraints from the first iteration.

What teams typically run into in the first month

The site doesn't oversell. Here's what's true about getting started:

  • The first ingest run is messy.Five years of unstructured notes don't synthesise cleanly the first time. The LLM produces some pages that need merging, some that need splitting. It gets better fast — typically by the second or third weekly run, the wiki settles into a coherent shape.
  • The five foundation files take 1–2 hours to write properly. Don't shortcut this. The foundations anchor every interpretation downstream; a vague baseline produces vague synthesis. An hour with the senior PMs and the Head of Product, before any evidence lands, pays for itself ten times over.
  • Reconcile is the discipline that pays off six months in, not week one. The first month's value is mostly in the wiki forming. Reconcile becomes load-bearing as decisions accumulate and delivery starts contradicting last quarter's assumptions.
  • Hosting decisions matter. Self-host first to learn the shape of your project; the open-source distribution runs on your laptop. Move to usewin.now once you know what you want operationally — same software, hosted for you with daily snapshots and a 30-day export window if you ever change your mind.

The methodology earns trust slowly. The first week is mostly setup; the first month is mostly the wiki forming; the second quarter is when the compound interest starts showing up.

Where to go next