Three steps: gather your win data, identify common patterns, encode them as criteria with weights. First rubric should take an hour; subsequent refinement happens over weeks.
Step 1 — Gather win data
Export your last 20–50 closed-won customers from CRM with firmographic and technographic enrichment. If you have fewer than 20 customers, use your best prospects-in-pipeline as proxies. Note customer-fit quality from CS perspective: who's renewing, who's expanding, who's churning.
Step 2 — Identify the patterns
Across your wins, what's true of most of them? Common patterns: similar size band, similar industries, specific technographic markers (using competitor X, on tool Y), geographic concentration, similar buyer titles. These are your starting criteria.
Step 3 — Encode as a rubric
Open ICP → New Rubric. For each pattern you identified, add a criterion: criterion type (firmographic, technographic, behavioral), value or range, weight (1–5). Add 2–3 disqualifiers for segments you've confirmed don't work (too small, regulated industry you don't serve, specific competitors).
Step 4 — Test against known data
Click Preview → the rubric scores every account in your Golden DB. Look at the top 100 — do they look like your customer base? Look at the bottom — are you correctly excluding accounts that wouldn't be a fit? Both ends should feel right; iterate criteria until they do.
Step 5 — Save and deploy
Save the rubric. Assign it to an agent or use it for list-building. Plan to revisit in 30 days with actual deployment data.
