$ man clay-wiki/closed-won-enrichment

Closed-won enrichment takes your best existing customers and reverse-engineers what made them a fit. Instead of guessing at ICP criteria, you look at the companies that already bought and extract the patterns. What industries are they in? What employee count range? What tech stack? What signals were present before they closed? You enrich your closed-won accounts in Clay, build a scoring model from the patterns, and then apply that model to score new prospects. Your best customers become the template for finding the next ones.

PATTERN

Most teams build scoring models from theory. "We think companies with 500+ employees in SaaS are a fit." That's a guess. The closed-won approach builds scoring from evidence. 1. Pull all closed-won accounts from HubSpot or Salesforce 2. Import into a Clay account table 3. Enrich every account with full firmographic data 4. Look at the data — what do the winners have in common? 5. Build a scoring model based on actual patterns 5 points per data point is the default. If 80% of your closed-won accounts have 500+ employees, that gets 5 points. If 90% are in SaaS or FinTech, that gets 5 points. If 60% use Salesforce, that gets 3 points (weaker signal). The score must be self-explanatory at a glance. Use Clay's scoring integration, not formulas, for handoff readability.

PATTERN

Go deeper than standard firmographics: • Employee count and revenue — the basics • Industry and sub-industry — look for clusters • Tech stack — which tools appear most often? (BuiltWith, Wappalyzer) • Funding stage and last raise — are your customers mostly Series B-C? • Geography — regional patterns? • Growth signals — hiring velocity, office expansion • Engagement history — what content did they engage with before closing? • Sales cycle length — how long from first touch to close? • Champion profile — what title/role drove the deal internally? The more data points you enrich, the more patterns emerge. Most teams stop at industry and employee count. That's the minimum. The closed-won play rewards depth.

PATTERN

Once you've enriched your closed-won accounts and identified patterns, the lookalike model writes itself: 1. Export the common attributes (industry clusters, employee range, tech stack patterns, geography) 2. Build a company qualification prompt that scores new prospects against those attributes 3. Weight the signals by frequency in your closed-won data (80% frequency = 5 points, 60% = 3 points, 40% = 2 points) 4. Set thresholds: 8-10 = high-fit lookalike, 6-7 = medium-fit, below 6 = not a match 5. Apply the model to your TAM list or inbound leads The lookalike model isn't static. Re-run closed-won enrichment quarterly. As you win new customers, the patterns evolve. The model should evolve with them.

PRO TIP

Closed-won enrichment is the prerequisite for the Alumni Effect. You can't track alumni if you don't know who your closed-won contacts are. The closed-won table feeds two downstream workflows: 1. Lookalike scoring — applied to new prospects 2. Alumni tracking — monitoring job changes from closed-won contacts Both workflows start with the same data. Enrich once, use twice.

ANTI-PATTERN

Using total customer count instead of closed-won. Not every customer is a good customer. Filter for your best deals — highest contract value, fastest close, lowest churn. Build the scoring model from your wins, not your entire customer list. Another mistake: building the scoring model once and never updating it. Your ICP evolves. Your product evolves. Re-enrich and recalibrate every quarter. The market shifts. Your scoring model should shift with it.