Custom Scoring (ICP, Churn, Upsell)
Last updated: April 22, 2026
Custom Scoring (ICP, Churn, Upsell)
Score any account or contact on a custom dimension with weighted inputs. CRED's scoring engine lets commercial and RevOps teams build ICP fit, churn risk, or upsell propensity scores using any field or signal in the graph, with transparent explanations.
Data strength at a glance
Rating scale: Strong — dense, fresh, validated. Solid — good coverage with known gaps. Partial — core fields only. Thin — custom work needed. None — not covered.
| Industry | Overall rating | What it means here |
| Technology / SaaS | Strong | Rich input feature set. |
| Consumer Brands, CPG & Retail | Strong | — |
| Professional Services | Solid | — |
| Financial Services | Solid | — |
| Sports, Media, Talent | Solid | Industry-specific fields available. |
| Healthcare, Manufacturing, Real Estate | Partial | Input field depth limits score quality. |
Core data — coverage, accuracy, recency
Values below are indicative ranges for mid-market+ accounts in developed markets. Ask your CRED account team for live coverage on your specific list, ICP or region.
Score templates
| Field | Coverage | Accuracy | Recency | Notes |
| ICP fit score (template) | Strong | Benchmarked on wins | On-demand | DS model |
| Churn risk (template) | Solid | Requires historical labels | On-demand | DS model |
| Upsell propensity (template) | Solid | Requires historical labels | On-demand | DS model |
| Custom weighted-field score | Strong | — | On-demand | Fully configurable |
Explainability
| Field | Coverage | Accuracy | Recency | Notes |
| Per-field contribution breakdown | Strong | — | On-demand | — |
| Peer percentile on each input | Strong | — | On-demand | — |
| What changed since last score | Strong | — | On-demand | Diff explanation |
Signals we fire for this use case
| Signal | How it's detected | Typical latency |
| Score moved above / below threshold | DS model | Daily |
| Input field decayed (affects score) | DS + decay detector | Daily |
| New input field added to model | Model updates | On-demand |
What this use case does not cover
- Guaranteed causal interpretation (scores are predictive, not causal)
- Replacement for a proper experimentation / A/B programme
Related use cases
- Prospecting & ICP fit
- CRM hygiene & sync
- Hiring & org-change alerts
Have a specific question about whether CRED can answer this for your team? Contact your CRED account team for a tailored review of coverage against your list or ICP.