TL;DR: Each record has an icp_score and persona_score, each 0–100. On the lead / account detail page, click either score to see exactly which parameters contributed and how much.
Where scores appear
- Records list (
Leads,Accounts) — two columns:ICP scoreandPersona score. Sortable. - Lead detail page — the
Qualificationpanel near the top shows ICP code + score, Persona code + score, and the Stage chip (Cool / Warm / In-market / Hot, computed from engagement, not ICP / Persona). - Audience composition — when you expand an audience, the breakdown shows score distributions across members.
Score range and meaning
- 0–25 — qualified but weakly aligned. Record matched Filters and dodged Excludes, but matched few or no Score parameters.
- 26–50 — partial match. Several Score parameters matched but the high-weight ones didn't.
- 51–75 — solid match. Most Score parameters matched; some high-weight ones missed.
- 76–100 — strong match. Top-weighted parameters matched. These are your priority records.
The mapping is calibrated per workspace. After a few weeks of outcome data, the Insights page reveals where reply rate jumps — usually around the 60-mark — and you can adjust your audience thresholds accordingly.
Per-parameter breakdown
Click the score on the lead detail page. A drawer opens showing each Score parameter as a row:
Parameter | Mode | Value on record | Weight | Contribution
Uses Snowflake | Score | yes | 5 | +5 / 5
Uses Okta | Score | unknown | 3 | 0 / 3
Funding stage | Score | Series B | 4 | +4 / 4
Hired a CISO in 12mo | Score | yes | 8 | +8 / 8
**Subtotal** | | | 20 | **17 / 20**
Industry (Filter) | Filter | software | — | passed
Employee count (Filter) | Filter | 240 | — | passed
Competitor (Exclude) | Exclude | no | — | passed
17 / 20 normalizes to 85 / 100 on the ICP score field.
The drawer shows Filter and Exclude parameters too, even though they don't contribute to score — useful for confirming the record actually passed them.
When the score "doesn't match expectations"
Common causes:
- The record has missing enrichment data for a high-weight Score parameter. "Uses Snowflake" is null on the record → 0 contribution from a weight-5 parameter → score 5 points lower than it should be. Enrich the record and re-score.
- A Filter is failing silently. The record is showing up in the wrong ICP because your most specific ICP's Filter excluded it, so it fell back to a broader ICP. Click into the Qualification panel and check the active ICP code.
- The Persona is unfit but the ICP looks good. Score parameters on the Persona don't carry over to the ICP score. A 90 ICP / 30 Persona record is a "right company, wrong contact" — common, and usually points at needing a different Persona definition.
Re-scoring a record
Three ways to trigger a re-score:
- Automatic on enrichment. When new enrichment data lands, the relevant ICP / Persona re-scores automatically.
- Manual on the lead detail page. Click the overflow menu →
Re-qualify this lead. The score recomputes against the current Live ICPs and Personas. - Bulk from the Leads page. Select rows →
Bulk actions→Re-qualify. Useful after editing an ICP / Persona definition mid-stream.
What scores do NOT include
- Engagement. Engagement score is a separate field tracking signal recency and source weight. A record can have a high ICP score (right company) but low engagement (no recent activity). The Stage chip reflects engagement, not ICP fit.
- Outcome history. Reply rate, meeting rate, signup rate are tracked at the ICP / Persona aggregate level and surface in Insights — they don't change an individual record's score.
- Manual overrides. There's intentionally no "boost this lead" button. If you want a specific lead surfaced, add a parameter that captures its trait and re-score the workspace.