What is lead scoring?
Lead scoring ranks leads or accounts by how closely they match the ideal customer profile and how recently they have shown relevant activity. A score is a prioritisation of what to work first — not a prediction that a particular deal will close.
Why it matters
A rep can work a fixed number of accounts in a day. Scoring decides which ones, and doing it explicitly is the difference between a rule the whole team can see and argue with, and each rep’s private instinct about which names look promising.
What goes into a score
- Fit: how closely the account's firmographics match the ideal customer profile.
- Role fit: whether the contact holds a title and seniority you actually sell to.
- Activity: what the account has done recently, and how recently it did it.
- Signals: dated events at the account, weighted by how much each type matters to your category.
- History: the accounts your team has worked, won, and explicitly dismissed.
Three ways to set the weights
| Model type | How the weights are set | When it fits |
|---|---|---|
| Rule-based | By hand, by the revenue team | Early, when there is little closed-won history to learn from |
| Fitted to outcomes | Derived from which accounts actually closed | Once there is enough history for the pattern to be real |
| Hybrid | Fitted weights, with hand-set rules and hard disqualifiers on top | Most teams, most of the time |
Common mistakes
- Reading a score as a probability. It ranks; it does not forecast.
- Scoring on activity alone, which promotes researchers, students and competitors to the top of the queue.
- Running a model with no disqualifiers, so an account you cannot sell to can still score well on engagement.
- Never recalibrating. A score fitted to last year’s product prioritises last year’s market.