Most lead scoring advice is written for companies with a marketing operations team, a data warehouse, and thousands of leads a month to train a model on. If you're a founder or a three-person sales team, that advice is worse than useless, it's a distraction. You don't need a 40-variable predictive model. You need a simple, honest way to decide which leads to call first, and the discipline to actually use it. Here's a lead scoring model you can build in a spreadsheet this afternoon.
Why small teams need scoring at all
When you only have a handful of hours a day for outreach, the cost of working the wrong leads is enormous. Chase ten poor-fit prospects and you've burned a day for nothing; spend that same day on ten strong-fit, high-intent leads and you might book three meetings. Scoring isn't bureaucracy, it's how a small team makes its scarce selling time count. The goal is a single number that tells you: call this one before that one.
The two things that actually predict a good lead
Strip away the complexity and almost all useful lead scoring reduces to two dimensions:
- Fit — does this company look like the customers who already buy, stay, and get value? Fit is about firmographics: size, industry, tools they use, geography.
- Intent — is there a signal that they're actively in-market or experiencing the problem you solve right now? Intent is about timing: a trigger event, a website visit, a reply, a hiring signal.
A point system you can run in a spreadsheet
Assign points for fit and intent separately, then combine. Keep the criteria few and concrete so scoring takes seconds per lead:
| Category | Signal | Points |
|---|---|---|
| Fit | Company size in your ICP band | +20 |
| Fit | Industry matches your best customers | +15 |
| Fit | Uses a tool that signals fit (right CRM, stack) | +15 |
| Fit | Wrong size or industry | -20 |
| Intent | Replied or asked a question | +25 |
| Intent | Visited pricing or booked-a-call page | +20 |
| Intent | Recent trigger event (funding, new hire in relevant role) | +15 |
| Intent | No activity in 30+ days | -10 |
- Hot (call today): 50+ points, strong on both fit and intent.
- Warm (this week / nurture): 25-49 points.
- Cold (ignore or long-term drip): under 25.
Fit is stable, intent is fresh
The most common small-team scoring mistake is treating the score as permanent. Fit rarely changes, a 120-person logistics company is a 120-person logistics company next month. But intent decays fast. A prospect who replied last week is hot; the same prospect silent for a month has cooled. Rescore intent regularly (a weekly pass is enough for most small teams) while leaving fit alone unless the company genuinely changes.
How to calibrate without a data team
You don't need statistical rigor to make this accurate, you need a feedback loop. Once a month:
- Pull the last 20-30 deals you won and score them retroactively. If your best customers aren't scoring hot, your fit criteria are wrong, fix the point weights.
- Pull deals you lost or that ghosted after a demo. If they scored hot, you have a false-positive problem, usually an intent signal that's noisier than you thought.
- Adjust two or three weights, not the whole model. Small, frequent corrections beat a big annual overhaul.
When to let software do the scoring
A spreadsheet works until volume outgrows it. Once you're evaluating hundreds of new prospects a week, manually scoring each on fit and intent becomes the bottleneck. That's the point to move scoring into tooling that applies the same criteria automatically. [AutoReach](/register) scores lead quality with AI as it finds businesses, using the same fit-and-intent logic, and it learns from which leads you approve or reject so the scoring sharpens toward your real ICP over time, the automated version of the monthly calibration loop above.
If you'd rather own the scoring model outright, including the reply and outcome data it learns from, you can fine-tune an open-weight model on your accept/reject history with a platform like InfoPlatform.ai instead of relying on a vendor's built-in scoring.