Five weighted areas
- Revenue follow-up and pipeline risk: 40 points
- Duplicate integrity: 25 points
- Bad records and CRM usability: 20 points
- Ownership and routing: 10 points
- General organization: 5 points
Transparent scoring
LeadRecoverly CRM Health Score Methodology
The score is an explainable prioritization tool built from visible findings in the current scan. It is not a statistical guarantee, revenue forecast, or substitute for human review.
Direct answer: The implemented score starts at 100 and deducts points for current findings across five weighted areas: revenue follow-up and pipeline risk (40), duplicate integrity (25), bad records and usability (20), ownership and routing (10), and general organization (5).
Explainable deductions
Each active deduction includes a finding key, visible title, count, importance, and recommended action. A scan with no actionable findings scores 100.
Finding examples
Implemented categories include duplicate contacts and companies, bad records, missing owners, incomplete contacts, lifecycle gaps, stale deals, past-close deals, unowned deals, missed follow-up, attribution gaps, cleanup queues, and recovery priority queues.
Record-level priority scores
Some queues also use 0–100 review scores based on current record signals. Thresholds determine whether a record qualifies and whether it is displayed as a review, warning, or critical item.
Configured value estimates
Dollar figures are directional estimates used to prioritize review. They depend on user-configured assumptions and CRM values and are not guaranteed recovered revenue.
Sample versus full scans
The free plan is a sample scan limited to 500 contacts, 250 companies, and 250 deals. Paid-plan limits are larger, so a sample score may not represent every record in a portal.
Comparison over time
Recover and Optimize can save score history. Monitoring compares completed scans and configured thresholds; it does not treat every score movement as a confirmed business outcome.
What the score does not prove
The score does not prove that a duplicate should be merged, a deal is recoverable, a contact is valuable, a routing policy is correct, or an estimate will become revenue.
Human review remains required
LeadRecoverly is read-only by default. Paid changes require the correct plan, Edit Mode, workspace role, preview, confirmation, relevant HubSpot scope, server-side validation, and action logging.
Continue reviewing the implementation
Implementation details and review guidance
How the deductions are intended to be used
The five category weights cap how much each class of issue can affect the overall score. Counts and severity influence visible deductions, but the score is designed to direct attention rather than rank companies against an external benchmark. Users should open the deductions, inspect the underlying records, and decide which findings are actionable in their own process.
Why the free score can differ from a paid scan
The free plan evaluates a bounded sample of contacts, companies, and deals. A sample can overrepresent or underrepresent a problem depending on record order and portal composition. Paid plans permit larger coverage, so the resulting score may change even when no records were edited. That difference is a coverage effect, not evidence that the earlier score was fabricated.
How estimates remain separate from facts
Record counts, field values, owners, dates, and activity signals come from available CRM data. A configured dollar estimate is a prioritization layer placed on top of those facts. The interface and documentation should keep those categories distinct so a user can tell what HubSpot supplied, what LeadRecoverly calculated, and what assumption the user configured.
Changes that do not automatically improve the score
A record edit only helps if the next completed scan no longer finds the qualifying condition. Dismissing a finding, exporting a row, assigning a task, or viewing a report does not by itself prove the CRM data improved. The scoring process should remain based on current scan evidence rather than UI activity.