Lead scoring model examples: three templates for acquisition
Most lead scores add up clicks until a number looks big. These three models separate who fits from who is active, forget old activity, and tell a rep what to do next.
Most lead scores are built the same way. Every form fill, email open, and page view adds points, and when the total crosses a line the lead goes to sales. A year later the top of the list is full of students, competitors, and people who opened a newsletter forty times, and reps have stopped looking at the number.
The fix is not a better point total. It is a model that answers two separate questions, forgets what went stale, and ties each result to an action. Below are three templates for the most common acquisition goals, which are booking demos, turning product trials into customers, and finding accounts where a buying group is forming. Take the weights as a starting point and tune them to your own closed-won data. Every record, count, and score on this page is demo data.
Why the old point total stopped working
Three things changed.
Buyers decide before they talk to you. In 6sense’s 2025 Buyer Experience Report, the winning vendor was already on the buyer’s Day One shortlist 95% of the time, and buyers made first contact about 61% of the way through their journey. The research that shapes the decision happens before a form fill. A score that only wakes up at the form is late.
Email opens are no longer a signal. Litmus reports that over 50% of email opens happen on a device with Apple’s Mail Privacy Protection turned on, which loads tracking pixels whether or not a person reads the email. Any model that still gives real weight to opens is scoring a privacy setting.
The response is too slow to waste on the wrong people. When Workato submitted demo requests to 114 B2B companies, the average email reply took nearly 12 hours and almost 20% never replied by email at all. Reps are not lazy. They are spread across too many records. Salesforce’s State of Sales research found reps spend 60% of their time on non-selling tasks, so the 40% left has to go to the right leads first.
Two scores, not one
Every model below produces two scores, each out of 10 to one decimal, and keeps them apart.
- Fit asks whether this is the kind of person and company you win. It comes from company and contact properties, and it changes slowly. A 7.0 and up is a strong fit, 4.0 to 6.9 is a decent fit, and under 4.0 is a weak fit.
- Engagement asks whether they are active right now. It comes from behavior, and it decays. A 6.0 and up is active now, 3.0 to 5.9 shows some recent activity, and under 3.0 is quiet.
Put together, every lead lands in one of nine cells, from fit 7.0 and up with engagement 6.0 and up (strong fit, active now) to fit under 4.0 with engagement under 3.0 (weak fit, quiet). HubSpot’s own lead scoring tool also keeps fit and engagement as separate scores, so the models map onto fields your team already has.
Keeping the two apart is the point. A single blended number hides the most important cases. A perfect-fit account that went quiet and a poor-fit student who downloaded six ebooks can end up with the same total. As two scores, they are obviously different leads that need different plays.
Five rules every model follows
- Weights add up to 100 within each score. That keeps the scale readable and forces a trade-off every time you add a factor.
- Engagement decays. A pricing-page visit from last week matters. One from last spring does not. Each engagement factor below lists how quickly it should lose value. HubSpot supports this directly, from monthly to yearly decay.
- Opens score zero. Clicks, replies, and visits carry the email signal instead.
- Disqualifiers subtract. Competitors, personal email domains, students, and existing customers should pull a score down or route the record somewhere else, not quietly add up to a 7.0.
- The reasons are visible. A rep should be able to open a record and see which factors made the score. A score nobody can explain gets ignored.
Model 1: Book more demos
For sales-led teams that want reps talking to the right inbound and near-inbound leads first.
Fit score
| Factor | Weight | What counts as a strong match |
|---|---|---|
| Company size | 30% | Employee band where your best customers sit |
| Industry | 25% | The sub-verticals you win most often |
| Role and seniority | 25% | Titles that appear on closed-won deals as buyer or champion |
| Required system in place | 20% | They run the CRM, platform, or edition your product needs |
| Disqualifiers | Subtract | Competitor domain, personal email, student, current customer |
Engagement score
| Factor | Weight | Decays |
|---|---|---|
| Visits to pricing, comparison, or integration pages | 35% | Half every 30 days |
| Bottom-of-funnel content, such as case studies, ROI tools, and comparison guides | 20% | Half every 60 days |
| Webinar or event attendance, not just registration | 15% | Half every 60 days |
| Three or more site visits in 14 days | 15% | Half every 30 days |
| Email clicks and replies | 10% | Half every 30 days |
| Other people from the same company active this month | 5% | Half every 30 days |
The biggest change from older templates is that form fills alone are not a factor. The demo request is the outcome you are trying to predict and speed up, so the model looks at the research that comes before it. When a demo request does come in, it should route on the spot regardless of score. The score decides the order, not whether someone gets a reply.
- Company size 30% 8.7
- Industry 25% 8.2
- Role and seniority 25% 7.6
- Required system in place 20% 5.9
The math, with demo data. 8.7 at 30% is 2.61, 8.2 at 25% is 2.05, 7.6 at 25% is 1.9, and 5.9 at 20% is 1.18, which adds to 7.74, or a fit score of 7.7.
This lead is a strong fit at 7.7, carried by company size and industry. The weakest line is the required system, where enrichment could not confirm which CRM edition they run. That is the question for the first call, and the rep can see it before dialing.
Model 2: Turn product trials into customers
For product-led or hybrid teams, where people sign up on their own and the question is which trials deserve a human.
Fit uses the same factors as Model 1, with company size weighted a little higher, since trial volume brings in many individuals and very small teams. Engagement changes completely. What people do inside the product predicts buying far better than what they do on your website.
Engagement score
| Factor | Weight | Decays |
|---|---|---|
| Reached the activation event, the action most paying customers take early | 35% | Half every 30 days |
| Invited a teammate | 25% | Half every 60 days |
| Connected an integration or imported real data | 20% | Half every 60 days |
| Active on three or more separate days in the trial | 15% | Half every 14 days |
| Visited pricing or plan comparison while in the trial | 5% | Half every 14 days |
Define the activation event from your own data. Look at paying customers and find the action most of them took in their first days that non-payers skipped. That one action usually matters more than the rest of the list combined.
A teammate invite is the product version of a buying group forming. One person trying a tool is curiosity. Two people from the same company is a project.
- Reached the activation event 35% 9.1
- Invited a teammate 25% 7.4
- Connected an integration or imported real data 20% 6.3
- Active on three or more separate days 15% 8.2
- Visited pricing or plan comparison 5% 4.7
The math, with demo data. 9.1 at 35% is 3.185, 7.4 at 25% is 1.85, 6.3 at 20% is 1.26, 8.2 at 15% is 1.23, and 4.7 at 5% is 0.235, which adds to 7.76, or an engagement score of 7.8.
At 7.8 this trial is active now. The activation event and the repeat days carry it, the teammate invite is a week old and already decaying, and the one weak line is pricing, which they have barely looked at. That tells the rep to lead with what the team is building, not with the plan comparison.
Model 3: Find accounts where a buying group is forming
For teams selling larger deals, where the question is not which lead is most active but which account is moving. Gartner puts B2B buying groups at five to 16 people across as many as four functions. Scoring each of them as a separate lead scatters one buying decision across five or ten records and several owners.
Score the company instead. Fit comes from company properties, the firmographic and technographic factors in Model 1 without the role line. Engagement rolls up from every contact at the account.
Engagement score
| Factor | Weight | Decays |
|---|---|---|
| Distinct buying roles engaged, such as economic buyer, champion, and technical evaluator | 30% | Half every 30 days |
| Number of engaged contacts at the account | 25% | Half every 30 days |
| High-intent page visits across all contacts | 25% | Half every 30 days |
| Timing signal, such as a new leader in a buying role, hiring for the team you serve, or funding | 20% | Half every 90 days |
Roles carry more weight than raw headcount on purpose. Three SDRs reading your blog is less meaningful than one sales leader and one CRM admin looking at pricing and integrations in the same week. The second pattern is what a real evaluation looks like.
- Distinct buying roles engaged 30% 6.8
- Engaged contacts at the account 25% 7.3
- High-intent page visits across all contacts 25% 8.4
- Timing signal 20% 3.6
The math, with demo data. 6.8 at 30% is 2.04, 7.3 at 25% is 1.825, 8.4 at 25% is 2.1, and 3.6 at 20% is 0.72, which adds to 6.685, or an engagement score of 6.7.
At 6.7 the account clears the 6.0 line, mostly on pricing visits from two of the four contacts and a champion who has shown up. The timing line is low because the only trigger is a job post from two months ago, already half decayed. If a new sales leader lands next month, that line jumps and the account moves up the list on the next sync.
In HubSpot, fit, engagement, and combined scores all work on company records as well as contacts, so the account score can live on the company and the contact scores can stay on the people.
What to do with each score band
A score is only useful if each pair of bands has a play. Here is a starting point for all three models. The cutoffs are a starting point too. Set your own where conversion clearly rises in your past leads.
| Engagement 6.0 and up | Engagement 3.0 to 5.9 | Engagement under 3.0 | |
|---|---|---|---|
| Fit 7.0 and up | Rep outreach the same day. Named owner, personal note that references what they looked at | Rep follows up this week with a specific reason to talk | Keep them on the shortlist with targeted ads, useful content, and an invite to something small. Alert the owner the moment engagement climbs |
| Fit 4.0 to 6.9 | Qualify quickly. One short call or a self-serve path to the answer they are after | Nurture with content matched to what they read, and rescore in 30 days | Nurture only. No rep time yet |
| Fit under 4.0 | The trap. Lots of activity, poor deals. Check whether the fit data is wrong, then send to self-serve, partners, or a lighter-touch offer | Self-serve and content only | Leave them alone |
Two notes on the table. First, the cell with fit under 4.0 and engagement 6.0 and up deserves a second look every quarter. Sometimes it is noise, and sometimes it is a new segment your fit model has not caught up to. Second, bands are cutoffs on the numbers, not labels stamped on a record. As volume changes, move the cutoffs so the top-left cell holds roughly as many leads as your reps can work properly in a week.
Then automate the plays. A workflow enrolls records with fit 7.0 and up and engagement 6.0 and up, and alerts the owner. An ads audience builds from records with fit 7.0 and up and engagement under 3.0. A record that climbs from a 2.4 engagement to a 6.3 after a pricing visit gets picked up on the next sync, without anyone re-sorting a list.
Check that the model works
Every quarter, pull every lead created that quarter and look at conversion by band.
- Leads with fit 7.0 and up and engagement 6.0 and up should convert at a clearly higher rate than everything else. If they do not, the weights are wrong.
- Look at closed-won deals that started with fit under 7.0. Each one points at a factor the model is missing or underweighting.
- Look at leads with fit 7.0 and up and engagement 6.0 and up that went nowhere. Each one points at a factor that is overweighted, or a disqualifier you have not written down yet.
Adjust one or two weights at a time, write down why, and check again next quarter. A model someone set by hand is a model someone can explain, and that is what keeps reps trusting it.
Where BoostIdeal fits
All three models are plain weighted factors, so you can run them in a spreadsheet to start. BoostIdeal runs them on every record, working with HubSpot or a CSV you upload. You build the model and set the weights, scores out of 10 write back to HubSpot with their reasons, and it can score one record from another, so a contact can carry its account’s buying-group score. It also runs several models side by side, which is how a team keeps the demo model, the trial model, and the account model going at once.
What you end up with
Not a bigger number. Two scores out of 10 that say who fits and who is active, engagement that forgets what went stale, no weight on signals a privacy setting can fake, and a play for every pair of bands. Pick the model that matches your motion, start with these weights, and let your own wins and losses tune them from there.