Building a Lead Scoring Model That Sales Actually Trusts
Most lead scoring models die the same way: marketing builds one based on assumptions about what a good lead looks like, sales ignores the score within a month because it doesn't match what they're seeing on calls, and everyone goes back to working leads in the order they arrived.
A model sales actually trusts is built backwards from what closed, checked against sales's own read of the leads, and revisited often enough to stay accurate. Here's the order that tends to work.
Vendors Covered in this Article
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
How do you build a lead scoring model from closed deals?
Pull the last two or three quarters of closed-won and closed-lost deals, and for each one, gather the firmographic details and the behavior that led up to the sale: company size, role of the buyer, pages visited, content downloaded, how they first found you. Don't start from a list of things that sound like they should matter; start from what the data actually shows in common across the deals that closed.
This step alone often overturns assumptions a team has held for years. A title that felt like an obvious buyer signal sometimes shows up just as often in the lost deals as the won ones, while a smaller, less obvious signal, like visiting the pricing page twice in one session, turns out to correlate strongly with closing.
Score fit and behavior as two separate numbers
Fit is about whether this is the kind of company and person who could realistically buy: company size, industry, role. Behavior is about what they're actually doing right now: pages visited, emails opened, content downloaded. Blending these into one score hides useful information, because a perfect-fit company that's barely engaged looks the same on paper as a mediocre-fit company that's deeply engaged.
Keep them as two numbers, or at least two visible components of one score, so sales can see at a glance whether a lead is a strong fit who isn't engaged yet, or a strong engagement signal from a company that might not be the right size. Those two situations call for very different follow up.
How do you set a lead score threshold from your own close rate?
A common mistake is picking a threshold that just feels reasonable, like the top score band out of a hundred point scale, without checking what that threshold actually predicts. Instead, sort your historical leads by score and find the point where close rate genuinely jumps, the natural breakpoint in your own data rather than an arbitrary round number.
That breakpoint becomes your qualified threshold. It won't be a clean number, and that's fine; a threshold derived from your actual close rate will hold up better under scrutiny from sales than one chosen because it looked tidy on a slide.
Get sales to sign off before you turn it on
Walk a sample of scored leads past the reps who'll be working them, and ask directly whether the score matches their own read of each one. Where it doesn't match, dig into why: is the model missing a signal reps consider important, or is the rep's instinct actually wrong and just hasn't been tested against real close data yet.
This conversation matters more than any technical refinement to the model. A scoring model sales helped shape gets used; one that just appears in the CRM one Monday morning gets ignored within a few weeks, no matter how statistically sound it is.
Recalibrate every quarter, not once and done
Your close rate by score band will drift as your product, market, and ideal customer shift, and a model built on last year's deals quietly becomes less accurate every quarter it goes unchecked. Rerun the same close-rate-by-score analysis quarterly, and adjust the weights or the threshold when the data has clearly moved.
Treat the first version of the model as a starting point, not a finished product. Teams that build a model once and never touch it again are usually the same teams complaining a year later that sales stopped trusting the score.
Follow this order when you build the model:
- Pull two or three quarters of closed won and closed lost deals, and note the firmographic details and buying behavior each one showed before the outcome.
- Score fit and behavior as two separate numbers, so a perfect fit company that is barely engaged does not look the same as a mediocre fit that is very active.
- Sort historical leads by score and set the threshold where close rate genuinely jumps, instead of at a round number that only feels reasonable.
- Walk a sample of scored leads past the reps who will work them, and investigate wherever their own read disagrees with the score.
- Rerun the close rate by score band every quarter, and adjust the weights or the threshold once the data has clearly moved.
What Good Looks Like
A trustworthy lead scoring model has a threshold derived from your own close rate data, not a round number, and gets rechecked against that close rate every quarter.
Building The Capability (5-Stage Skill Ladder)
How to Get Started
Disclosure: We may earn a commission if you buy through some links on this page. It doesn't change what we recommend.
ClickUp is a reasonable place to track the sign-off conversations with sales and the quarterly recalibration schedule, so the model doesn't quietly go stale.
If email engagement is one of your behavior signals, a platform like ConvertKit gives you open and click data at the individual level instead of a blended list average.
Frequently Asked Questions
How many factors should go into a lead scoring model?
Fewer than most teams start with. A model with five or six factors that genuinely correlate with closing, checked against real deal data, beats one with twenty factors that sound reasonable but were never tested. Start small, confirm each factor actually predicts closing, and only add more once the core model is working.
What if sales and the data disagree about what makes a good lead?
Dig into the specific disagreement rather than picking a side by default. Sometimes reps are picking up on a real signal the model is missing, like tone on a call, that's worth finding a proxy for. Other times the instinct doesn't hold up against actual close data, and that's a useful, if uncomfortable, thing to surface together.
Should behavior score decay over time if a lead goes quiet?
Yes, generally. A lead who was highly engaged two months ago but hasn't opened an email or visited the site since isn't the same as one who engaged yesterday. Building in decay, so behavior score naturally fades without new activity, keeps the score reflecting current interest rather than a snapshot from months back.
Who should own the lead scoring model, marketing or sales?
Marketing usually owns building and maintaining the model, but sales needs a real voice in defining a qualified lead. The underlying data and tooling typically sit in marketing's systems, and sales should flag when the score stops matching reality. A model owned entirely by one side tends to lose the other's trust over time.
About the numbers
This guide doesn't quote a sourced benchmark. Figures in it are estimates or general guidance, so check them against your own numbers.
Related Guides
Which Conversion Benchmarks Actually Apply to Your Funnel
A guide to reading conversion rate benchmarks correctly: which published numbers are close enough to compare against, and which ones will mislead you.
Deciding Whether Community-Led Growth Fits Your Product
A decision guide for whether an owned community will actually drive growth for your product, and what it takes to run one instead of just launching one.
Choosing the Right Content Syndication Channels
A comparison of the main ways to syndicate content, from partner newsletters to paid distribution networks, and how to tell which one fits your goal.
A Checklist for Organic Search Visibility, Beyond Keywords
A practical checklist covering the technical, content, and authority work that decides whether your pages actually rank, plus the pitfalls that undo it.
Building an Executive Marketing Dashboard Leadership Actually Reads
A worksheet walkthrough for building an executive marketing dashboard with the handful of numbers leadership actually needs, not everything you track.
A Smarter Way to Win Back Users Who Went Quiet
A step-by-step approach to re-engaging users who stopped opening your product, from segmenting dormancy to knowing when to stop trying.