How to score LinkedIn leads by fit and intent (with a worked formula)

The formula, term by term

Bingle's lead score is:

score = 100 × fit × (0.3 + 0.7 × intent)

Fit runs from 0 to 1 and says who someone is: how well their job title, seniority, industry, location and company size match the ideal customer you set. Intent also runs from 0 to 1 and says what they did lately: viewed your profile, commented on a competitor's post, started a new role. Matching your ideal customer or being active on LinkedIn isn't intent on its own; intent is something dated that they or their company did.

Two choices do most of the work. Fit multiplies everything, so a lead with fit 0.5 can never score above 50 and a strong signal can't rescue the wrong person. And the 0.3 is a floor: a perfect fit with no signal still scores 30, because a great-fit person is worth knowing before the timing is right. Signals decide the other 70 points.

The score orders the list; Bingle's buckets say what to do with it. Ready to contact means a strong fit and a recent signal from the person themselves. A strong fit with no signal yet is Warm: worth watching, not worth a message today. And each Bingle source has a Minimum ICP fit you can raise or lower; people below it are hidden rather than ranked.

Signal strengths and how fast they fade

Each signal has a strength and a half-life: the time after which it counts half as much, then half again after the same time. So a signal's pull on intent at any moment is roughly its strength × 0.5 ^ (days since it happened ÷ half-life). These are the strengths and fade times Bingle publishes on its LinkedIn buying signals page:

Broadly, things people do to you (view your profile, react to your post) beat things they do near you (engage with a competitor or an influencer), which beat simply being active. Company events like hiring and funding last much longer. A profile view is about this week; a new role is about this quarter.

A profile view from two days ago still counts at about 82% of its strength (0.5 ^ (2 ÷ 7)); at two weeks a quarter, at a month about 5%. Without a fade, last quarter's views would crowd out this week's.

Combining several signals

Adding signals up breaks quickly: three weak ones would beat one strong one, and the total would run past 1. Bingle's rule is that signals stack with diminishing returns, so a second signal adds less than the first, and signals from colleagues at the same company add to the account at a discount.

If you're scoring by hand, a simple way to get diminishing returns is to treat each signal as independent evidence and combine them the way you'd combine probabilities:

intent = 1 − (1 − s1) × (1 − s2) × (1 − s3) ...

where s1, s2 and s3 are each signal's faded strength. Two signals of 0.6 give 1 − 0.4 × 0.4 = 0.84, not 1.2, and a third of 0.6 only takes it to 0.94. That's the method the spreadsheet below uses. It's a sound default, and you can tighten it by counting a repeat of the same signal, like a third like on a competitor's post, for less than a new kind of signal.

Three example leads, worked through

Three made-up leads for a seller of operations software in Australia, New Zealand and Singapore. The fit and intent values are illustrative, chosen to show how the formula behaves, not output from Bingle.

Lead A. Score = 100 × 0.85 × (0.3 + 0.7 × 0.75) = 100 × 0.85 × 0.825 = 70.1: a strong fit with a recent signal from the person, which is what Ready to contact means. A week later the view counts half as much; if intent falls to about 0.37, the score drops to 47.5. Still a good fit, no longer the top of the list.

Lead B. Say the new role, a month into its three-month fade, is worth 0.55 on its own and the competitor comment, five days into its two-week fade, is worth 0.55 too. Combined, intent = 1 − (1 − 0.55) × (1 − 0.55) = 1 − 0.45 × 0.45, or about 0.8. Score = 100 × 0.9 × (0.3 + 0.7 × 0.8) = 100 × 0.9 × 0.86 = 77.4. Neither signal alone matches A's view, but together, on a better fit, they put B first.

Lead C has the freshest signal and the lowest score: 100 × 0.5 × (0.3 + 0.7 × 0.9) = 46.5. A marketing coordinator at an agency looking at your profile is curiosity, not a buyer, and at a fit that low most people would set a source's Minimum ICP fit above them, which hides them in Bingle. If you'd still like to write to viewers like C, our guide on what to message someone who viewed your LinkedIn profile covers the wrong-person case.

How other tools describe their scoring

Every tool splits fit from intent somehow. They differ in how the two combine, whether signals fade and whether you can see why. Here's what each says on its own site, read 24 October 2026:

HubSpot's lead scoring comes with Marketing Hub or Sales Hub Professional or Enterprise; AI-built contact scores need Marketing Hub Enterprise. You give points to fit properties (region, company size, industry) and engagement events (page views, email clicks, forms, booked meetings). It's the better choice when your leads already live in HubSpot and the engagement you care about is email, website and meetings. Its help articles don't list LinkedIn activity, like a profile view, among the engagement examples.

Common Room pulls in signals from product usage, website, community and LinkedIn, and rolls person scores up to the account. In Settings, Scores you weight fit against behaviour and set bands from not a fit to excellent. If you sell a product with a free tier or an active community, that breadth beats a LinkedIn-only tool.

Gojiberry's FAQ says it checks each lead against your ideal customer and scores it, using signals such as profile visits, competitor engagement, job changes and hiring, but doesn't publish how they combine. A 22 October 2026 post on its blog describes filtering for non-negotiable fit first, then scoring behaviour. For the two side by side, see Bingle vs Gojiberry.

A spreadsheet version you can copy

For a list of up to a few hundred people, Google Sheets or Excel can run the same idea. It won't watch LinkedIn for you, so you'll log signals by hand; our weekly routine for finding leads who show buying signals covers where to look. Put headers in row 1, your fit weights in row 2 and leads from row 3:

1. Columns B to G: score each lead 0, 0.5 or 1 on job title, seniority, industry, company size, location and keywords against your ideal customer.
2. Row 2, B to G: your weight for each, adding up to 1. Title usually deserves the biggest share, because the role decides whether they can buy.
3. Column H, fit: =SUMPRODUCT(B3:G3,$B$2:$G$2).
4. Columns I to K for the first signal: its strength, the date it happened, its half-life in days. Columns L to N for a second signal. Pick strengths on a simple scale, say 0.9 for Strongest, 0.7 for Strong, 0.5 for Medium and 0.25 for Weak; these are round numbers for the sheet, not Bingle's settings. Take half-lives from the table above (7, 14, 21, 30, 45, 90 or 120 days).
5. Column O: =IF(I3="",0,I3*0.5^((TODAY()-J3)/K3)). Column P: the same for L, M and N.
6. Column Q, intent: =1-(1-O3)*(1-P3).
7. Column R, score: =ROUND(100*H3*(0.3+0.7*Q3),1). Sort by R, largest first, and set a minimum fit below which you don't bother.

Because it uses TODAY(), scores fall each morning as signals age. One simplification: the sheet treats both signals as independent, so if they're the same kind (two competitor likes), count the second for less, say half.

At a few hundred rows, the logging becomes the job. Bingle does the watching, scoring and ranking for LinkedIn signals and drafts each message for your OK; plans start at US$49 a month after a 7-day free trial with no card (pricing). For other options, see the best LinkedIn buying signal tools or compare tools side by side.

Questions

Sources

1. HubSpot Knowledge Base, Overview of the lead scoring tool. last updated 2 Sep 2026, read 24 Oct 2026
2. HubSpot Knowledge Base, Build lead scores to qualify contacts, companies, and deals. last updated 14 Aug 2026, read 24 Oct 2026
3. Common Room, AI Scoring. read 24 Oct 2026
4. Common Room, Follow up on the best leads fast with signal-based scoring. read 24 Oct 2026
5. Gojiberry, FAQ. read 24 Oct 2026
6. Gojiberry blog, Lead scoring best practices I wish I'd known sooner. published 22 Oct 2026, read 24 Oct 2026

More from the blog

• What to message a prospect who viewed your LinkedIn profile
• Best LinkedIn buying signal tools in 2026
• How to find people who engage with a competitor's LinkedIn posts
• How to know when a customer or prospect changes jobs (with or without Sales Navigator)
• What are the best SalesRobot alternatives for AI-run LinkedIn outreach?
• What are the best Dripify alternatives for LinkedIn outreach?

All posts