SANITIZED BUILD RECEIPT · builtgtm.ai/builds
Five steps turn "we have a product a lot of people use" into a ranked list of accounts with a named move on each. This page is the map. Every step opens to its own page in one click.
Who do we actually sell to - and which of our two motions are they in?
We collapsed a sprawl of conflicting definitions into one canonical ICP. The key call: in a product-led business, a self-serve user and a direct-sales buyer are two different customers with two different journeys. Telling them apart is what lets every step downstream treat them correctly.
Not every signal matters. We pick the ones that correlate with a win.
Anyone can collect signals. The work is knowing which ones predict a real B2B sales motion, then ignoring the rest. We tested each against won/lost history and kept the ones with lift: a prospect running a CRM in their stack is a 1.64× win signal. We look for proof they sell the way our customers sell - plus timing triggers like a new VP Sales, fresh funding, or sequence activation once they're in the product. The full list lives in the Signal Library.
Gather that evidence at scale, only on accounts worth a rep's time.
A provider waterfall pulls the correlated signals and the decision-maker: hiring and tech from Deepline, fit and playbook from Octave, the verified contact from FullEnrich. It stops as soon as it has a confident answer, and writes the result back onto the account so it's there when the rep opens it. We spend enrichment where the score says it'll pay off, not by blanket.
ICP qualifies and sets the base, product is the loud engine, buying signals are bounded boosts on top.
The enriched evidence becomes the score. Fit earns an account a real base; product usage does the heavy lifting because usage predicts the buy; each buying signal only ever adds points, capped - its absence is never a penalty. A deliverability gate and a hard-disqualify cap keep junk out. The same math runs across inbound, outbound, and product leads.
Final = min(100, base + engine + boosts) × email gate · Octave hard-disqualify caps at 40.
The score becomes context: what they use, what they don't, and the one move - for both an outbound prospect and a product user.
Sequences - 32 weekly users, 4 power users.
AI Compose, Smart Send, Meeting Copilot.
88-person sales team, hiring 4 SDRs, runs Salesforce, in-market today.
AI Compose - they hand-write every sequence. Contact: Jordan Lee (VP Sales).
Sequences + CRM sidebar - 24 weekly users, growing.
The whole AI suite (0 of 4).
Free workspace expanding fast; 2 people researching alternatives.
Lead with their own usage, pivot to the AI gap. First: enrich a sales DM.
