Lead scoring is one of those things every marketing team thinks it needs and very few implement well. The usual outcome is an elaborate model nobody trusts, sales ignores, and marketing quietly stops maintaining. The fix isn't a better algorithm — it's restraint. Here's how lead scoring really works and how to build one that survives contact with reality.
What lead scoring is for
Lead scoring is a system for ranking leads by how likely they are to become customers, so that sales spends time on the right people and marketing knows who's ready for a handoff. That's it. It's a prioritisation tool, not a crystal ball. Keeping that purpose front of mind is what stops a model from ballooning into something unusable.
The two dimensions: fit and behaviour
Fit (who they are)
Does this person match your ideal customer? Job title, seniority, company size, industry, geography. A senior decision-maker at a target-size company in your core industry scores high on fit regardless of what they've done.
Behaviour (what they do)
How engaged are they? Email opens and clicks, page visits (especially pricing and product pages), content downloads, demo requests, webinar attendance. Behaviour signals intent and timing.
The crucial insight is that these two dimensions are different questions. A perfect-fit lead who's done nothing isn't ready for sales. A highly active lead who's a terrible fit isn't worth a rep's time. Mature models keep fit and behaviour somewhat separate rather than blending them into one meaningless number.
How to actually build one
Start by defining 'qualified' with sales
Before assigning a single point, sit with sales and agree what a sales-ready lead looks like. If marketing and sales don't share that definition, no scoring model will fix the handoff — it'll just automate the disagreement.
Keep the model small
Resist the urge to score everything. Pick a handful of high-signal fit criteria and a handful of high-intent behaviours. A model with ten well-chosen rules beats one with fifty rules nobody can explain. You should be able to look at any lead's score and say why it's that number.
Add negative scoring
Just as important as adding points is removing them. Students, competitors, job seekers, free-email-domain signups, and people who've gone inactive for months should lose points. Negative scoring keeps the top of your list clean.
Decay over time
Behaviour scores should fade. Someone who downloaded a guide six months ago and vanished isn't as hot as someone who did it yesterday. Without decay, your model accumulates stale points and slowly loses meaning.
The pitfalls that make scoring useless
Over-complication
The number-one killer. The more elaborate the model, the less anyone understands it, the less they trust it, and the faster it rots. Complexity is not sophistication.
Building it in a vacuum
If marketing builds scoring without sales, sales won't use it. Co-ownership isn't optional — it's the whole point of the exercise.
Set and forget
A scoring model needs periodic review. Are the high scorers actually converting? If your best leads aren't closing, the model is wrong and needs recalibrating against real outcomes.
A simple model that works
Here's a sane starting point. Assign fit points for a handful of ideal-customer attributes. Assign behaviour points for a few high-intent actions, with the biggest weights on the strongest signals (pricing-page visits, demo requests). Apply negative points for clear disqualifiers. Set a threshold that, when crossed, notifies sales and creates a task. Then watch what converts and adjust. Start simple, prove it works, and only add complexity when a real gap demands it.
When you might not need scoring at all
One honest caveat: not every team needs a scoring model. If your lead volume is low enough that sales can review every lead by hand, a formal model may be premature overhead. Lead scoring earns its keep when volume exceeds what humans can triage manually. Don't build sophisticated machinery to solve a problem you don't yet have.
Want to see which platforms list lead scoring as a required skill? Browse current Marketing Ops and automation roles — it shows up constantly, and knowing how to do it simply is a genuine differentiator.