Lead scoring answers two questions in order. Does this prospect match the kind of customer we close? And are they showing signs of buying now? The first question is fit. The second is intent. A working lead-scoring system combines both into a single score that drives routing decisions.
Why lead scoring exists
Without scoring, every lead gets the same attention. With unlimited sales capacity that is fine. In reality, sales teams have constraints. An AE can run 15 to 25 discovery calls per week. An SDR can place 60 to 80 outbound touches per day. If the team treats a hot inbound demo request the same as a cold trial signup, the hot leads cool and the cold leads consume capacity.
The job of lead scoring is to triage. Hot leads route to the AE within an hour. Warm leads route to the SDR with an SLA. Cold leads route to nurture or to outbound prospecting. The score itself is just a number; the value is in what the number causes the team to do.
The two-dimensional model
Most working B2B lead scoring uses two dimensions:
- Fit: Does this prospect match our ICP? Firmographic signals only.
- Intent: Are they showing buying behavior? Behavioral signals only.
The two dimensions are independent. A prospect can be a perfect ICP fit who has not engaged at all. Or a poor ICP fit who is hitting the pricing page three times a week. Each combination implies a different action.
| Low intent | High intent | |
|---|---|---|
| Low fit | Disqualify | Light touch, check for misclassification |
| High fit | Outbound nurture | Hot lead, route to AE |
Dimension 1: Fit scoring
Fit is firmographic. It does not change based on what the prospect does. It changes only when the company changes.
The five fit attributes to score on:
- Company size. Headcount or revenue band. Most B2B SaaS has a 3x-wide ICP band (e.g., 50 to 500 employees). Score 0 for outside the band, 1 for inside.
- Industry. The verticals you have closed before. Score 0 for unverified vertical, 1 for verified, 2 for top-3 verified vertical.
- Geography. Where the buyer operates. Score 0 for outside go-to-market territory, 1 for inside.
- Tech stack. Tools the prospect uses that signal fit. CRM, sales engagement, data provider. Score 0 to 2 based on overlap.
- Role of the contact. The title that matches your buying-role list. Score 0 for outside, 1 for buying role, 2 for decision-maker.
A simple sum produces a fit score from 0 to 8. Working thresholds:
- 0 to 2: Poor fit. Disqualify.
- 3 to 5: Moderate fit. Worth nurture.
- 6 to 8: Strong fit. Worth direct outreach.
Dimension 2: Intent scoring
Intent is behavioral. It changes whenever the prospect does something.
The intent signals to track:
- Website visits. Pricing page visit = high signal. Blog visit = low signal. Career page visit = no signal (probably a job seeker).
- Email engagement. Opened the email = low signal. Clicked a link = medium signal. Replied = high signal.
- Trial or demo activity. Signed up for a trial = high signal. Used the product = very high signal.
- Content consumption. Downloaded a guide = medium signal. Watched a webinar = high signal.
- Sales engagement. Booked a meeting = highest signal. Accepted a connection = medium signal.
- Third-party intent. Bombora, G2 Buyer Intent surge alerts = medium signal. Indicates the company is researching your category.
Weight each signal by how predictive it is in your specific funnel. The signals worth most are the ones that show up in 80 to 90 percent of closed deals retrospectively.
A working intent score:
- Pricing page visit: 30 points.
- Demo request: 100 points.
- Trial signup: 80 points.
- Email reply: 50 points.
- Meeting attended: 200 points.
- Guide download: 10 points.
- Webinar attended: 25 points.
- Third-party intent surge: 20 points.
Total scores above 80 typically warrant immediate attention. Above 150, route to AE.
Combining the two dimensions
Multiply, do not add. A high-fit prospect with zero intent is not the same as a low-fit prospect with high intent.
The combined score:
combined = fit_score * intent_scoreFor a fit score of 6 and intent score of 80: combined = 480. For a fit score of 2 and intent score of 100: combined = 200.
The first prospect gets more attention even though the second prospect has higher absolute intent. The math is right: you want to spend capacity on prospects who both fit and engage.
Routing rules
The score's value is in what it triggers. The typical routing tier:
| Combined score | Tier | Routing | SLA |
|---|---|---|---|
| 400+ | Hot | Direct to AE | Under 1 hour |
| 150 to 399 | Warm | SDR for qualification | Under 4 hours |
| 50 to 149 | Cool | Nurture sequence + SDR follow-up | Under 24 hours |
| Below 50 | Cold | Marketing nurture only | Weekly |
The SLAs matter as much as the routing. A hot lead that sits for 24 hours is no longer hot. Speed of response is one of the strongest predictors of close in inbound B2B.
BANT, MEDDIC, SPICED: when to use which
Scoring helps prioritize. Qualification frames help disqualify. The two work together but they are not the same thing.
BANT (Budget, Authority, Need, Timing). Short-cycle deals under $25,000 ACV. Simple to apply, easy to teach. The qualification call confirms each of the four attributes.
MEDDIC (Metrics, Economic Buyer, Decision Criteria, Decision Process, Identify Pain, Champion). Enterprise deals over $100,000 ACV. Multi-stakeholder, long cycle, custom procurement. Requires multiple conversations to map all six attributes.
SPICED (Situation, Pain, Impact, Critical Event, Decision). Modern alternative to BANT that adds a critical event (a specific deadline or trigger forcing action). Better than BANT for mid-market because it surfaces the urgency dimension.
Pick one and use it consistently. Mixing frames across reps creates accountability gaps.
Where lead scoring goes wrong
Mistake 1: Scoring on activity that does not predict close. Logo visits, social-share counts, podcast plays. Score only on signals that retroactively predict closed deals in your historical data.
Mistake 2: Static scores that never decay. A prospect who visited the pricing page 6 months ago is not as hot today as one who visited yesterday. Build time decay into the intent score: half-life of 30 to 90 days.
Mistake 3: Adding rather than multiplying. Fit and intent are not interchangeable. A high-fit low-intent prospect is a different lead than a low-fit high-intent prospect. Multiplication preserves the distinction.
Mistake 4: Score thresholds that route to no one. If "warm leads route to SDR for qualification" but the SDR is at quota, warm leads pile up. Match the routing to the available capacity.
Mistake 5: Treating the score as the qualification. The score gets the prospect routed; the human still has to qualify. A high score is permission to invest sales time, not a guarantee of fit.
How to build a lead-scoring model from scratch
The 6-step build process:
- Pull the last 50 closed-won deals. Find what they had in common in firmographics and behavior.
- Pull the last 100 lost or disqualified leads. Find what was missing.
- Identify the 5 to 8 attributes that separate winners from losers. These become the scoring inputs.
- Assign weights. Higher weight to attributes that appear in 80%+ of wins. Lower weight to attributes that appear in 30 to 60%.
- Set routing thresholds. Pick numbers that produce roughly the right volume of hot leads for your team's capacity.
- Review monthly. The model drifts as the product, market, and team change. Revisit the weights every 30 to 60 days.
The first version will be imperfect. That is fine. A working scoring model improves with use; a perfect model that does not ship improves with nothing.