Separate fit from engagement
Two questions matter. Is this the kind of company or person we serve well, and are they showing interest now? Fit uses attributes such as industry, company size, role, geography, and technology in use. Engagement uses behavior such as page visits, content downloads, email responses, and event attendance. Keep the two scores distinct, since a perfect-fit company with no engagement and a highly engaged student are different cases needing different handling.
- Fit score from firmographic and role data.
- Engagement score from recent behavior.
- A combined view, such as a grid, to decide the response.
Build the model from your customers
Look at the customers you have won and kept, and identify what they share. Compare them with leads that never progressed. Use those differences to decide which attributes earn points. Include negative scoring for disqualifiers, such as competitors, students, or regions you do not serve. Start simple and adjust after reviewing results, since a first guess will not be perfect.
Behavior, recency and decay
Interest fades. A pricing page visit this week means more than a webinar attended a year ago. Apply recency weighting, so points decay over time and old activity does not keep a dormant lead on top. Give higher value to high-intent actions such as demo requests or pricing views than to casual reading. Be careful with signals that can be automated, like opens of email, which may reflect security software rather than interest.
Thresholds and the handoff to sales
Define when a lead becomes sales-ready, and what happens next. Agree the threshold with sales leaders, with a service level for follow-up, and route leads using clear rules. Leads below the threshold can enter nurture programs. Let sales reject leads with a reason, and feed rejections back into the model. Without agreement, marketing will optimize for volume and sales will ignore the output.
Lead routing and territory design →Data quality and enrichment
Scores are only as good as the data. Missing company size or inconsistent job titles degrade fit scoring. Use validation at capture, standardize titles into role groups, and consider enrichment sources to fill gaps, subject to your privacy obligations and consent. Remove duplicates and match leads to existing accounts so scoring reflects the whole buying group.
Data 360 →Measure and recalibrate
Track conversion from lead to opportunity and to customer by score band, time to follow-up, and rejection reasons. If high scorers do not convert better than low scorers, the model needs change. Review quarterly with marketing and sales together. Consider testing changes on a portion of leads before rolling them out. A model that is reviewed regularly stays useful, while one that is never touched quietly decays.
Making it stick
Document the model in plain language, including what earns and loses points, and share it with both teams. Appoint an owner. Keep the number of rules small enough that a new colleague can understand it in a few minutes. Celebrate wins that came from the model, and investigate misses without blame, since both teach something about the market.
Decisions to settle before configuration starts
- Definition of ready. Agree what makes a lead sales-ready, jointly with sales.
- Fit attributes. Choose the firmographic and role factors that matter.
- Behavior weights. Rank actions by intent, with decay rules.
- Service level. Set follow-up times and rejection handling.
A realistic first 90 days
- Days 1 to 30. Analyze won and lost leads, agree the definition of ready, and clean the data.
- Days 31 to 60. Build fit and engagement scores, set thresholds and routing, and pilot with one team.
- Days 61 to 90. Add decay and feedback loops, publish conversion by band, and adjust.
Pitfalls to avoid
- Mixing fit and interest. A single blended number hides what to do.
- No decay. Old activity keeps dead leads at the top.
- No sales agreement. Leads are ignored if sales did not help define them.
- Never reviewing. Markets and messaging change, so the model must too.
