What is Lead Scoring? Models, Examples and Thresholds
Lead scoring is a way to rank your leads by giving each one points: points for fit (how closely they match your ideal customer) and points for engagement (what they’ve done, like visiting pricing or requesting a demo). When a lead passes a threshold, it moves to the next stage, usually from marketing to sales. The aim is simple: your sales team spends its time on the leads most likely to buy.
Why lead scoring matters
Without lead scoring, sales either calls everyone, which wastes hours on students and competitors, or calls no one, and good leads go cold. Scoring makes the handoff between marketing and sales explicit.
Scoring also feeds your ad platforms. If you send “qualified” events back to Google or Meta, they optimize toward better leads, not just more of them.
How lead scoring works
A lead scoring model has two halves. Here is a simple B2B version for a company selling to SaaS firms with 20–500 employees.
| Fit criteria | Points | Engagement criteria | Points |
|---|---|---|---|
| Company size 20–500 | +20 | Requested a demo | +40 |
| Industry: SaaS or fintech | +15 | Visited pricing page (last 30 days) | +15 |
| Job title: Head of, VP, C-level | +15 | Opened 3+ emails (last 30 days) | +5 |
| Target country | +10 | Downloaded a guide | +10 |
| Free email domain (gmail.com) | −15 | Visited careers page | −10 |
| Competitor or student | −50 | No activity for 60 days | −20 (decay) |
Set a threshold based on history, for example fit ≥ 40 and total ≥ 70 to become a marketing-qualified lead. The MQL vs SQL definitions should be agreed with sales before you switch it on.
Example: scoring three leads
- Lead A: Head of Growth at a 120-person SaaS company in the US, visited pricing and requested a demo. Fit 20 + 15 + 15 + 10 = 60. Engagement 40 + 15 = 55. Total 115: route to sales today.
- Lead B: Marketing manager at a 5-person agency, downloaded two guides, opened six emails. Fit 0. Engagement 10 + 10 + 5 = 25. Total 25: keep nurturing.
- Lead C: VP Finance at a 300-person fintech, visited pricing once, using a Gmail address. Fit 20 + 15 + 15 + 10 − 15 = 45. Engagement 15. Total 60: high fit, low engagement, so a good target for a personal email or a LinkedIn Thought Leader Ad sequence.
Most small teams don't need a complicated score. I'd start with fit alone: is this the right company and the right role? Add two or three behavior signals, such as a pricing visit or a demo request, then sit down with sales once a month to check which scored leads actually became deals. A model nobody reviews quietly drifts until sales stops trusting it.
How to build a lead scoring model
-
Look at closed deals
Pull the last 20–50 won and lost deals. What did winners have in common: size, role, industry, actions?
-
Define fit
Turn the pattern into 4–6 fit criteria with points, including negatives for clear non-buyers.
-
Define engagement
Weight high-intent actions (demo, pricing, contact) far above low-intent ones (opens, blog visits).
-
Add decay
Reduce engagement points as activity ages, so last year's webinar doesn't keep a lead "hot".
-
Agree the threshold with sales
Pick the score that routes a lead to sales and the service level for follow-up.
-
Automate the routing
Use marketing automation to assign owners, alert sales and start nurture sequences.
-
Review quarterly
Compare scores with outcomes. If high scorers don't convert, change the weights.
Myth vs fact
- MythMore points for more activity always means a better lead.
- FactActivity without fit is noise. A student who reads every blog post can outscore a buyer who visited pricing once. Separate fit from engagement and penalize clear non-buyers.
Add intent data as an extra engagement signal once your basic model works.
Frequently asked questions
What is a good lead score threshold?
There's no universal number. Set it by looking at past leads that became customers and finding the score most of them reached. Start conservatively, then adjust after reviewing results with sales each quarter.
Should I use predictive lead scoring?
Predictive scoring uses machine learning on your CRM history. It needs enough closed deals to learn from. Most small B2B teams should start with a manual model they understand, then test predictive scoring as data grows.
What is lead score decay?
Decay reduces engagement points as actions get older, for example halving them every few months. It keeps scores reflecting current interest, so sales isn't chasing people who were active a year ago.
Who should own lead scoring?
Marketing usually builds and maintains the model, but sales must agree the criteria and threshold. Review it together, or the scores won't be trusted.