Systems · 8 min read
A practical lead-scoring model — sales will actually use
A score is useful only when it changes a decision. Build one that makes the next action clear, shows its evidence, and can be improved from real outcomes.
Score fit, intent, and readiness separately. A single opaque number hides the reason a lead should—or should not—receive attention.
Why lead scores lose credibility
Most failed scoring models share two traits: they reward easy-to-capture activity, and nobody can explain the number. A prospect who visits three blog posts might outrank a buying committee that fits perfectly but has not yet filled out a form.
The alternative is a transparent model that helps people make a small number of operational decisions: route now, nurture, research, or exclude.
Use three scores instead of one mystery score
| Score | What it answers | Example inputs |
|---|---|---|
| Fit | Is this the kind of account and buyer we serve well? | Segment, company size, geography, technology, role. |
| Intent | Is there credible evidence of a relevant need now? | High-intent page views, demo request, event attendance, verified external signal. |
| Readiness | Can the team take an appropriate next action? | Contactability, routing completeness, owner capacity, consent state. |
You can still use a combined priority score, but retain the components. “High priority because fit is strong and intent is recent” is a usable explanation. “86 points” is not.
Start with rules you can defend
Do not begin with a complex predictive model. Start with observable inputs and weights that sales, marketing, and RevOps can challenge constructively.
| Signal | Suggested treatment | Why |
|---|---|---|
| Target-account fit | Strong positive weight | It protects the team from pursuing structurally poor opportunities. |
| Demo or pricing intent | Strong positive weight with a short expiry | It often changes the right next action, but gets stale quickly. |
| Relevant product usage | Positive weight when tied to a conversion hypothesis | Usage can indicate need; raw activity alone may not. |
| Generic content activity | Low weight or no weight | It is easy to generate and easy to overinterpret. |
| Suppression, competitor, or disqualifying segment | Hard exclusion | Some rules should override every positive signal. |
Make score bands trigger actions
| Band | Meaning | Next action |
|---|---|---|
| Priority | Strong fit and timely intent; record is ready. | Route to an owner or a reviewed outbound task with evidence attached. |
| Research | Potentially valuable but missing a key fact. | Enrich, validate the signal, or identify the relevant person. |
| Nurture | Fit exists but intent is weak or stale. | Enter an appropriate educational or signal-monitoring path. |
| Exclude | Outside the motion or not safe to contact. | Suppress from activation and retain the reason. |
Include evidence in the handoff
A salesperson should never receive a scored lead with no context. Attach the three components, the contributing signals, their dates, and the recommended action.
- Fit: “200–500 employee SaaS account in the priority segment.”
- Intent: “Pricing page viewed yesterday after an integration guide.”
- Readiness: “Verified marketing leader, assigned territory, no suppression.”
- Action: “Respond to inbound request” or “review for a role-specific outbound touch.”
Review the model against outcomes
Every month, compare score bands with meetings held, accepted opportunities, conversion, sales feedback, and time-to-follow-up. Look for signals that create busywork, segments the team consistently rejects, and high-quality opportunities the model missed. Adjust the inputs and document the change.
The model is not the strategy. It is a shared, inspectable way to turn strategy into action.