What each report contains

Three reports, in order: diagnosis before context before action.

Every PMFKit run produces three reports. The order matters. Strategies that aren't grounded in observed signals are generic; PMFKit forces the chain to be visible. A founder who disagrees with the final verdict can walk back to a specific claim in Discoveries and challenge the underlying source, because every claim is sourced and labeled.

Report 1, Discoveries

The question it answers. What do we actually see on this product today?

Discoveries is built from a rendered HTML crawl of your public site (Playwright), a backup static parse, Schema.org and OpenGraph metadata, your robots.txt and sitemap, plus public domain age and a public traffic estimator where available.

What you get back: the brand identity we read off the page, the product summary, pricing and packaging, distribution links and CTAs, traffic order of magnitude, and traction signals (GitHub stars, ProductHunt presence, social mentions). Every claim carries a trust pill: Verified, Likely, or Unknown. See the trust pills page for what each label means.

If the crawl fails (a hard paywall, a JavaScript only page that won't hydrate, a robots disallow), the report says so explicitly rather than degrading silently. The trust pill goes Unknown and the source breadcrumb is one click away.

Report 2, Context

The question it answers. How does this compare to the rest of the market?

Context is built on top of Discoveries. We pull search rank positions for your product name and category queries, run embedding based competitor similarity to surface the top five adjacent products, query GitHub for repo presence and activity signals, and pull public social signals (X, ProductHunt, Indie Hackers, Hacker News) for mention counts.

What you get back: a competitor map with similarity scores, a value proposition clarity reading versus alternatives, the geographic scope of your footprint (what country specific signals we picked up), and the strategic posture of nearby products (what they emphasize, what they charge, what channels they use).

Context is where the scope of your market gets sized. A small public footprint isn't a problem on its own; it's a signal about the stage you're at, which feeds Strategies.

Report 3, Strategies

The question it answers. What should this team do next?

Strategies inherits from Reports 1 and 2 plus our internal channel success stories knowledge base and a stage classifier (Idea Validation, Early Traction, PMF, Growth). It's the only report that recommends.

What you get back: the stage diagnosis (what stage are you actually at, not what stage you'd like to be at), the single focus area for the next 30 to 90 days, three to five milestones to validate that focus, distribution channel recommendations grounded in comparable products, and the Continue, Pause, or Kill verdict with a confidence label.

See the verdict page for what each verdict means and how to read the confidence label.

Reading the reports together

The reports are designed to be read top to bottom. If you disagree with the verdict in Strategies, scroll up to the Discoveries claim it rests on. If a Discoveries claim is wrong, the trust pill tells you whether we read it directly (Verified) or inferred it (Likely or Unknown). The full evidence chain is visible by design.