Market Research

Scan a market in a fixed order, evidence every claim, and hand back a digest — not a strategy essay.

Scan order

1. Competitor discovery — direct search patterns ("<category> software", "<category> tools 2026"), comparison/review sites (G2, Capterra, AlternativeTo), launch platforms (Product Hunt, BetaList, Show HN), and "<competitor> alternative" / "<competitor> vs" queries to surface who else gets compared against whom.
2. Pricing-page teardown — capture every tier verbatim: name, price, billing period, currency, seat/usage limits, feature gates. Don't normalize currency or round numbers; record what the page shows, note the annual-vs-monthly toggle discount if present.
3. Positioning read — pull the vendor's own headline, tagline, and top feature callouts. Label these as vendor claims, not verified fact — a pricing page describes intent, not measured outcomes.
4. Demand signals — search-volume proxies (cite the tool used, e.g. Google Trends direction, a keyword tool's estimate), community threads (Reddit/HN/niche Discord/Slack search for the pain, quote representative complaints with links), job postings (a company hiring for a role tied to the problem signals budget exists).

ICP hypothesis

For each candidate ICP state: who (role, company size, vertical), pain (the specific job-to-be-done), current workaround (spreadsheet, manual process, a competitor, nothing). Treat it as a hypothesis until at least one source — a community thread, a job post, or a competitor's own positioning — backs it. An ICP guess with zero backing sources is a gap, not a finding.

Evidence rules