Every lead score is the output of a model — a set of rules or logic that decides how promising a lead is. Get the model right and your reps open a list already sorted by who's worth calling first. Get it wrong and the score is noise: a number nobody trusts, ignored within a week. This is a tour of the models teams actually use, what each is good for, and how to assemble one that earns its place.
What is a lead scoring model?
A lead scoring model turns what you know about a lead into a single value that represents how likely it is to convert. The inputs fall into two big families: fit (how well the account matches your ideal customer) and intent (how much the lead is behaving like a buyer). A good model weighs both, subtracts points for poor-fit signals, and rolls the total up to a grade — like A–D — that a rep can read in a glance.
In one line: a scoring model is just “fit + intent − poor-fit signals,” tuned to your business and expressed as a grade.
The five models you'll actually meet
There are dozens of named scoring methodologies, but almost all of them are a combination of five building blocks. Understand these and you can read — or build — any model.
Firmographic / demographic (fit)
Points for how well the account matches your ideal customer — company size, industry, region, job title. Scores fit, independent of behavior, so a good-fit account rises even before it engages.
Behavioral / engagement (intent)
Points for the buying signals a lead shows — a demo request, a pricing-page visit, repeated returns. Scores interest, the clearest sign someone is actively in-market.
Fit-and-intent matrix
Combine the two on a grid: high-fit + high-intent is your A, high-fit + low-intent needs nurturing, low-fit + high-intent is a careful maybe. Most A–D grade systems are a one-dimensional version of this.
Negative scoring
Subtract points for poor-fit traits — a free-mail address, a competitor, a student title, an unsubscribe. The quiet half of a good model: it keeps the top grades meaningful.
Predictive / AI scoring
A model infers a score from patterns in historical data. Powerful at scale, but opaque — when a rep asks why a lead is hot, there's no rule to point to, and the model drifts on data you can't see.
Rule-based vs. predictive: which should you use?
This is the real decision, and it's less about sophistication than about trust and control. Predictive (or “AI”) scoring infers a score from patterns in your historical data. It can surface signals a human wouldn't spot — but it's a black box. When a rep asks why a lead is an A, there's no rule to point to; when the model drifts because the underlying data shifted, you may not notice for a quarter; and it needs a large, clean history to work at all.
Rule-based scoring takes the opposite stance. You write the rules, so every point traces back to a decision you made. Identical inputs always produce the same score. Change a rule and you can see exactly what moved and why. The trade-off is that you do the thinking — but that's also the point: reps trust a grade they can read, and you can tune the model as your ideal customer changes instead of waiting for an algorithm to relearn.
Our take: for most teams, an explainable rule-based model you actually maintain beats an opaque model you don't trust. That's why KudosCRM scores with transparent rules by design — not a predictive black box.
How to build a simple model that works
You don't need a complex model to get value — you need one you understand and revisit. A reliable starting point:
- 1
Define your ideal customer in one sentence
Company size, industry, and the role you sell to. This becomes your firmographic fit rules.
- 2
Pick your one strongest intent signal
Usually a demo or pricing request. Give it the most weight — it's the clearest 'in-market' sign.
- 3
Add two or three fit rules
Points for target company size and industry; a smaller bump for the right job title.
- 4
Add one or two negative rules
Subtract for free-mail domains, competitors, or students so they don't crowd the top grades.
- 5
Set your grade thresholds
A common split is A at 80+, B at 60+, C at 40+, D below — adjust once you see real leads land.
- 6
Watch, then tune
Let real leads grade for a week or two, sanity-check a few against your gut, and adjust the weights. Iterate, don't over-engineer.
How KudosCRM does lead scoring
KudosCRM implements the rule-based approach end to end. You build rules across four signal categories — intent, firmographic, behavioral, and negative — each adding or subtracting points toward a 0–100 score that rolls up to an A–D grade. Every lead carries a per-rule breakdown, so the grade is always explainable. Change a rule and the whole base re-grades in one pass. It scores on the signals you feed it through forms, the API, and automation — and for a step-by-step setup, see how to set up lead scoring rules.
One honest note: automatic page-visit and email-engagement tracking is on the roadmap rather than live, so behavioral rules score the signals you already capture today. That's the same transparency principle applied to the product itself — we'd rather show you exactly what fires than imply tracking we don't do.