Blog · Sales and revenue

How to Build a Lead Scoring Model: A Simple Point System for Small Businesses

The short answer

A lead scoring model gives each lead points for fit (are they the kind of customer you win: right industry, size, location, and role?) and intent (are they showing interest: asking for a quote, visiting your pricing page, replying to emails?). Subtract points for signs they'll never buy.

Base the points on your last 50 to 100 won and lost deals, not guesswork. Set a threshold, such as 70 out of 100, for leads that go straight to sales for a call within the hour, put the middle band into a follow-up sequence, and review the model every quarter by checking that higher-scoring leads really do close more often.

Fit vs. intent

TypeWhat it measuresExamples
FitWhether the lead looks like your best customersIndustry, company size, location you serve, job title, budget range
IntentWhether the lead is interested nowQuote or demo request, pricing page visits, email replies, repeat visits, event attendance

You need both. A perfect-fit company that never engages isn't ready, and a very active lead who's a student or a competitor will never buy. Scoring each separately also tells sales why a lead scored high.

A sample 100-point scoring model

This is a starting template for a B2B service business. Change the criteria and points to match what your own won deals have in common.

CriteriaPoints
Fit (up to 50)
Industry you serve well+15
Company size in your sweet spot+15
Decision-maker or budget holder+15
In a location you serve+5
Intent (up to 50)
Requested a quote, demo, or call+30
Visited your pricing page+10
Replied to an email or message+5
Attended a webinar or downloaded a guide+5
Negative points
No activity in 30 days−10
Unsubscribed from email−20
Student, job seeker, or vendor pitch−50
Competitor−100

Cap the score at 100 so a lead can't pile up points just by opening emails. Let intent points fade over time; a pricing page visit last week means more than one six months ago.

What to do at each score

ScoreLabelAction
70–100HotSend to sales immediately; call within the hour
40–69WarmPersonal follow-up within a day, then a short email sequence
0–39ColdNewsletter and helpful content until their score rises
Below 0DisqualifiedRemove from sales follow-up

The point of scoring is speed where it counts. Hot leads lose value fast; see how fast to follow up with a new lead.

How to build your model in 7 steps

  1. Pull your last 50 to 100 decided deals. Won and lost, with where each lead came from and what they did before buying.
  2. Find what winners have in common. Compare won with lost deals by industry, size, role, source, and actions taken. AI can help spot patterns in an export, but check what it finds against what your team knows.
  3. Pick 4 to 6 fit criteria and 4 to 6 intent actions. Keep it simple enough that sales understands and trusts it.
  4. Assign points by how strongly each predicts a sale. The actions most common among won deals get the most points.
  5. Add negative points and decay. Rule out leads who will never buy, and lower scores for inactivity.
  6. Set thresholds with sales. Agree on what score means "call now," and what sales must do with those leads and how fast.
  7. Build it in your CRM and test it on past leads. Most CRMs and marketing tools support point-based scoring. Score last quarter's leads and see whether the hot ones closed. Clean data first: duplicates and blank fields break scoring. This CRM cleanup checklist helps.

How to tell if your scoring works

Compare close rate by score band every quarter. A model that works shows a clear step up:

Score bandLeadsCustomersLead-to-customer rate
Hot (70–100)401025%
Warm (40–69)12097.5%
Cold (0–39)24031.25%

In this example, hot leads are 10% of volume but close 20 times as often as cold leads, so the team knows where to spend its time. If the bands close at similar rates, your points aren't measuring what matters; go back to your won deals and adjust. How to calculate close rate covers the formulas.

Rule-based vs. AI (predictive) scoring

Many CRMs offer AI or predictive lead scoring that learns from your past deals automatically. It can work well, but it usually needs a good amount of history, often hundreds of closed deals with consistent data, before its scores are reliable. Most small businesses should start with a simple point system they understand, then try predictive scoring once they have the data and can compare the two.

AI also helps around the model: researching a new lead's company to fill in fit fields, summarizing their activity for the salesperson, and drafting the first follow-up. See how to use AI for sales prospecting. Better scoring raises win rate, which means you need less pipeline coverage to hit the same target.

Frequently asked questions

What is lead scoring?

Lead scoring assigns points to each lead based on how well they fit your ideal customer and how much interest they are showing, so your team can contact the most promising leads first. Points can also be subtracted for signs a lead will not buy, such as being a competitor or unsubscribing.

How do I create a lead scoring model?

Pull your last 50 to 100 won and lost deals, find what the won deals have in common, choose four to six fit criteria and four to six intent actions, assign points by how strongly each predicts a sale, add negative points and decay for inactivity, agree on thresholds with sales, and test the model on past leads.

What is a good lead score?

It depends on your model. In a 100-point model, many businesses treat 70 or above as a hot lead ready for sales, 40 to 69 as warm, and below 40 as cold. The right threshold is the score above which leads close noticeably more often in your own data.

What is the difference between fit and intent in lead scoring?

Fit measures whether a lead looks like your best customers, such as industry, company size, location, and role. Intent measures whether the lead is interested now, such as requesting a quote, visiting the pricing page, or replying to emails. Good models use both.

Should a small business use AI lead scoring?

Predictive AI scoring usually needs a lot of consistent history, often hundreds of closed deals, to be reliable. Most small businesses should start with a simple point-based model and consider predictive scoring once they have enough clean data to compare the two.

How often should I update my lead scoring model?

Review it every quarter by comparing close rates across score bands. If high-scoring leads do not close noticeably more often than low-scoring ones, adjust the criteria and points based on your recent won and lost deals.

Spend your sales time on the right leads

Lead handling, follow-up speed, and funnel conversion are areas we review in every RevOps audit, along with pricing and the manual work AI can take on. The free Profit Check shows where your costs and margin stand against healthy ranges for your business type in about 3 minutes.

The Profit Snapshot ($2,500, about 2 weeks) finds the biggest quick wins. The AI Operations & Profit Audit ($7,500, about 3–4 weeks) covers your whole operation with a 30/60/90-day plan ranked by margin impact.

Take the Free Profit Check See audit options

This article is general guidance. The sample model and thresholds are starting points; your own won and lost deals are the best guide to what predicts a sale.