Product Discovery Consultant: Restart the Evidence Engine

Four habits: a standing interview cadence with recruiting that does not depend on heroics; an evidence template so findings accumulate instead of evaporating in notes; a synthesis rhythm that turns interviews into roadmap-ready claims; and the trace, so every major item cites its evidence or admits it has none.

The install matters more than the audit. Most companies buying discovery help have had good research done before; what they lack is the machine that keeps producing it. An engagement that ends with a research report changed nothing; one that ends with next month's interviews already scheduled changed the system.

Through the evidence layers the roadmap rule consumes: customer evidence that the problem is real, business evidence that solving it moves a metric, capacity evidence of honest cost. Discovery produces the first layer continuously; without it, the roadmap rule arbitrates between opinions dressed as items.

This is why the two pages cross-reference: a roadmap reset without a discovery engine starves within a quarter, and discovery without a roadmap rule produces insight nobody consumes. In embedded engagements I install them together, engine first, rule second, because the rule's first cycle needs evidence to chew.

Lighter and more continuous than the early days: a handful of conversations weekly, tied to live roadmap questions rather than open-ended exploration; usage data pairing every interview theme; and assumption tests sized to the decision, a landing page test for a bet, five interviews for a feature, not ceremony for either.

The published onboarding case shows the shape: the team did not need a research quarter, it needed the stall point located, users interviewed at that point, and one fix validated. Completion went from 52% to 81%. Discovery at this stage is targeted evidence for live decisions, which is why it fits inside 25 hours a month of leadership.

Product, structurally: whoever owns the roadmap owns the evidence it runs on. The consultant's exit deliverable is that ownership named, with the cadence on someone's calendar and the templates in the team's hands. Discovery owned by a research function alone produces reports; owned by product, it produces decisions.

The handover rule mirrors every capability engagement on this site: the fix without an owner decays within a year. Ask any discovery consultant, me included, what the last week of the engagement transfers, and expect a calendar and a habit, not a deck.

Scoped resets price inside the published monthly band, $2,000 to $10,000; embedded work runs at my flat $4,950 with the published hours band. The return is measured in misses avoided: features that failed the evidence test before they consumed a quarter, which is the cheapest failure a roadmap can have.

Run the return math on your last shipped miss: the engineering months it consumed, priced against five interviews that would have killed it in week one. Most companies find one such item on the current roadmap during the trace test, which alone covers a reset's fee.