The classic lead scoring model looked like this: +10 for VP title, +5 for company over 100 people, +5 for US or Europe, +20 for downloading a whitepaper, anything over 40 is a qualified lead. It was simple, it was easy to explain to a CRO, and it was wrong often enough that most SDRs ignored it. In 2025 the better teams have quietly replaced it, and the results are visible.
Why static scoring broke
Three reasons:
- Titles lie. A "VP of Growth" at a 10-person startup and a "VP of Growth" at a 5,000-person enterprise are not comparable buyers, but static scoring treats them identically.
- Intent is temporal. A CFO who was hot to buy in Q3 is cold in Q1. Static scores don't decay.
- The signals that matter now are behavioural. Did they visit your pricing page three times in a week? Did they share your content internally? Did they open four emails in two days? These are the things that actually predict a meeting.
The 2025 behavioural scoring model
A well-built 2025 model weights five things:
- Engagement recency. What did they do in the last 7 days? Last 30? Older activity decays fast.
- Engagement depth. A homepage view is 1 point. A pricing page view is 5. A demo video watched to completion is 15. A pricing page visit within 24 hours of opening an email is 25.
- Company-level signals. Job changes in the target persona, funding events, hiring signals, tech stack additions — these multiply individual scores.
- Negative signals. Unsubscribes, "not now" replies, and competitor mentions pull the score down hard.
- Decay. Every signal loses 50% of its weight after 30 days and 90% after 90 days. A hot lead from Q4 is a cool lead in Q1.
What this changes for SDRs
Under static scoring, every lead above 40 went into the outbound pool and got the same treatment. Under behavioural scoring, SDRs get a ranked daily list: these five prospects are showing buying intent right now, these 30 are warming, these 200 are cold. The day starts with the five, not the 235.
The tooling question
Most CRMs still ship with static scoring built in. Behavioural scoring usually lives in one of three places in 2025:
- A dedicated product (6sense, Common Room, Vector)
- A built-in feature of your sales engagement tool
- A custom SQL pipeline reading CRM + web analytics + sequence data
For SMB teams, option 2 is usually the right answer. For mid-market teams with real analytics staff, option 3 often wins. Option 1 is the default for enterprise but is overkill for most teams under 20 reps.
The honesty check
Behavioural scoring is only as good as the behaviour you can see. If your marketing team doesn't track known visitors, if your sequencer doesn't feed engagement back to CRM, if your reply handling is manual — the model has nothing to work with. Before you buy a scoring product, make sure the raw signals are actually flowing.
"We spent four months buying a predictive scoring tool. Then we spent two weeks realizing we had to build the data pipes first. The pipes mattered more than the model."
The five-row spreadsheet
If you want to start simple, make a spreadsheet with five columns: recency, depth, company signals, negative signals, decay. Fill it in manually for 50 prospects. See if it matches your gut about who is likely to reply. If it does, turn it into code. If it doesn't, fix the weights. This is how most 2025 scoring models actually start — not as a product purchase but as a spreadsheet that earned its keep.