If your ICP rubric is matching too few accounts ("low coverage"), three likely causes: criteria too tight, data gaps preventing matches, or wrong dimensions selected. Diagnose with the coverage panel.
What "low coverage" means
If your ICP matches fewer than ~10% of accounts in your Golden DB (when you'd expect more), or fewer than the TAM size you've estimated, coverage is low. The agent can only work matched accounts; low coverage means the agent has nothing to do.
Cause 1 — Criteria too tight
Too many criteria, or each criterion too narrow. If your rubric has 10 criteria all of which must roughly match, you've intersected your way to a tiny universe. Fix: reduce criteria count to 5–7; loosen ranges (size band 100–1000 instead of 100–200); make some criteria weighted rather than required.
Cause 2 — Data gaps
Your criteria require data that your enrichment doesn't have. If you require "uses Salesforce" but your technographic provider doesn't detect Salesforce well, you have low matches not because few accounts use Salesforce but because few have detected Salesforce. Fix: review per-field fill rate in Data Quality; add enrichment providers for fields with low fill.
Cause 3 — Wrong dimensions
Your criteria capture characteristics that don't actually predict fit. "Series B+" might not be your differentiator if your customers are mostly bootstrapped or PE-backed. Fix: re-examine your closed-won data; identify the actual common attributes (probably not what you initially assumed); rebuild criteria around the actual patterns.
When low coverage is correct
Sometimes your TAM really is small. If you're selling something specialized to a tight market, low coverage reflects reality. Calibrate by counting your hypothetical TAM (estimated count of fit accounts globally) and seeing if your rubric matches a reasonable fraction of that count.
