B2B Lead Scoring Model(Free Download)
A 0-100 B2B lead scoring model with point values for fit, behavior and intent, a 30-day decay rule, and 70/40 routing tiers — in Excel or your CRM.
Template preview
Lead scorecard

What's included
A 0-100 B2B lead scoring model with point values for fit, behavior and intent, a 30-day decay rule, and 70/40 routing tiers — in Excel or your CRM.
Template specs
Updated 2026-10-15
No credit card · Free forever
A B2B lead scoring model is an agreed point system that turns two separate questions — is this the right kind of company, and is this person actually buying — into one number a rep can sort by. This one runs on a 0-100 scale and splits the points into three buckets: up to 40 for firmographic fit (company size, industry, revenue), up to 45 for behavior and intent (pricing page visits, resource downloads, demo requests), and an engagement band for email and content activity, with a recency rule that subtracts points after 30 days of silence so a perfect-fit account that went quiet does not outrank a smaller one asking for a demo this week. Fit is scored once; behavior can fire again and again, which is why the behavioral ceiling is capped rather than open-ended — otherwise a newsletter regular outscores a VP of Sales. Threshold bands come pre-configured (70+ routes to sales, 40–69 enters nurture, under 40 stays in marketing) and every threshold is a single cell edit in the Excel scorecard. Negative rules belong inside the model rather than in a reject pile, so the steps below cover the deductions worth adding — unsubscribes, hard bounces, competitor and free-mail domains, careers-page visits, titles outside the buying group — and a lead that disqualifies itself then falls below the sales tier instead of sitting just under it. The weights that ship here are starting defaults; the steps and FAQs below show how to replace them with close rates from your own won and lost deals, which is the difference between a score reps argue with and one they quietly ignore. The CSV import maps each scoring rule to a HelloGrowthCRM automation, which means the same scorecard your team debates in a spreadsheet becomes the live routing logic in the CRM without re-keying anything.
The Problem This Solves
Sales teams waste time chasing low-quality leads while high-intent prospects go uncontacted for days, because nothing in the CRM separates the two. The usual failure is not the absence of scoring but a model nobody derived from data: weights set in a workshop, no deductions, no decay — so a subscriber who opens every email outranks a VP of Sales who booked a demo. Reps notice inside a month, stop trusting the number, and go back to working the list in arrival order, which is the problem the model was meant to fix.
Who It's For
B2B marketing, sales, and RevOps teams who get enough inbound to need an order of work — roughly 50 leads a month and up — plus anyone inheriting a scoring model that has stopped predicting anything. Below that volume the FAQs cover the stripped-down version worth running instead.
Use Cases
- Rank inbound leads so reps work the list in conversion order instead of arrival order
- Agree one MQL threshold that marketing and sales have both signed off on
- Route by score and territory: sales tier to a named rep, nurture tier to a sequence
- Replace workshop-set weights with close rates pulled from your own won and lost deals
- Add deduction and decay rules so disqualified and stale leads drop below the sales tier
- Score the account rather than the contact when several people from one company engage
- Audit a model that stopped predicting, by comparing the score spread of won and lost deals
How to Use This Template
- 1
Download the Excel scorecard, or import the CSV so each scoring rule lands as a HelloGrowthCRM automation
- 2
Write down your ICP first: the company size, industry, and revenue band your won deals actually share — not the one on the pitch deck
- 3
Calculate your baseline conversion rate (leads that became customers ÷ total leads) so every weight has a number to beat
- 4
For each candidate signal, pull the close rate of only the leads that carried it. Signals clearly ahead of the baseline earn points in proportion to the gap; signals level with it earn zero and come out of the model
- 5
Score fit once and behavior repeatedly, then cap the behavioral ceiling — without a cap, repeat activity outranks genuine fit
- 6
Add the deductions: unsubscribes, hard bounces, competitor and free-mail domains, careers-page visits, titles outside the buying group, and countries you do not sell into
- 7
Set the decay window (default: 30 days of no activity) and size the reversal against the points the activity earned
- 8
Set the tiers and the response time attached to each — sales tier contacted the same day, nurture tier into a sequence — because a score with no follow-up rule changes nothing
- 9
Recalibrate quarterly: plot won and lost deals by handoff score, check how many leads cleared the threshold and what share of those closed, then move the weights before you move the cutoff
Inside the template: sample data
The exact rows that ship with the B2B Lead Scoring Model download — replace them with your own records after import.
| Signal | Category | Points | Example |
|---|---|---|---|
| Company Size 50-200 | Firmographic | +15 | Mid-market ICP fit |
| Company Size 200+ | Firmographic | +20 | Enterprise ICP fit |
| Visited Pricing Page | Behavioral | +20 | High purchase intent |
| Downloaded Resource | Behavioral | +10 | Content engagement |
| Opened 3+ Emails | Engagement | +10 | Active nurture |
| Requested Demo | Intent | +30 | Direct sales signal |
| No Activity 30d | Decay | -15 | Engagement decay |
Customization guide
- Every point value — they are starting defaults, and your own close rates should overwrite them
- The tier cutoffs (70 for sales, 40 for nurture) and the names you give each tier
- The decay window: shorten it for a fast cycle, lengthen it for a long one
- The signal list itself — delete the signals your CRM does not reliably capture
- The firmographic bands, so the sizes and industries match the deals you actually win
- The 0-100 ceiling — keep it fixed, or this quarter's scores stop being comparable with last quarter's
- The split between the fit bucket and the behavior bucket; merging them hides whether a lead is a bad fit or just quiet
- The cap on behavioral points, which is the only thing stopping repeat activity from outranking fit
- The presence of at least one deduction rule, without which the model drifts upward until every lead looks warm
- Roll contact scores up to the account and weight seniority, so a forming buying group clears the threshold even when no single contact does
- Keep a separate baseline per lead source — paid search and referral rarely convert at the same rate, and one blended baseline flatters the weaker channel
- Back-test any weight change against last quarter's closed deals before you let it route live traffic
- Store the score the lead held at handoff, not just the current score; recalibration needs the number sales acted on
- Pair the score with a response time per tier, and review the leads that cleared the threshold but were never contacted
Frequently Asked Questions
Lead scoring measures engagement and intent, which is behavioral and can move week to week. Lead grading measures fit against your ICP, which is firmographic and rarely changes. This template adds both into one 0-100 number, but keep the two readable separately, because the pairing is what tells you the next action: high engagement with poor fit is a content reader, strong fit with no engagement is an outbound target, and only strong-on-both belongs in the sales tier. Most arguments about lead quality turn out to be an argument about that pairing.
Three families, plus a fourth that subtracts. Fit signals describe the company and the person — company size, industry, revenue band, job title and seniority. Behavior and intent signals describe what they did on your property: pricing page visits, resource downloads, webinar attendance, demo requests, trial sign-ups. Engagement signals describe the relationship over time — email opens and clicks, replies, repeat visits. Disqualifiers take points away. One test decides whether a signal is worth adding: can you see it in the CRM for most of your leads? A signal you only capture on one record in ten adds noise to the other nine.
Derive them rather than debate them. Start with your baseline conversion rate — leads that became customers divided by total leads. Then, for each signal, pull the close rate of just the leads that carried it. A signal sitting well above the baseline earns points roughly in proportion to that gap; a signal level with the baseline earns zero and should leave the model entirely, because it is telling you nothing you did not already know. The weights in this scorecard are defaults so you can run the model the same afternoon — treat them as a placeholder for that calculation, not a replacement for it.
Treat the first number as a hypothesis, not a setting. The default here is 70 for the sales tier and 40 for nurture. To find yours, score the last two or three quarters of closed deals retroactively and look for the band where deals that closed start clearly outnumbering the ones that did not — that band is your cutoff. Then watch two numbers for a quarter: how many leads clear the threshold, and what share of those close. Plenty clearing and few closing means the threshold is too low; sales reporting an empty queue while deals keep arriving from unscored leads means it is too high.
Anything that says this lead is not a buyer. Unsubscribes, hard bounces, email domains belonging to a competitor, free-mail addresses on a form meant for companies, careers-page and jobs-page visits, students and job seekers, titles outside the buying group, and countries you do not sell into. Size each deduction by how common that trait was among leads that went nowhere. A model with no deductions drifts until almost everyone sits in the seventies, and a score that cannot say no is not routing anything.
The model subtracts points after 30 days with no activity, so a score reflects current interest rather than a high-water mark from last quarter. The simplest way to extend it is to mirror the points: whatever an action earned, the reversal or the silence takes back. Match the window to your sales cycle — a 30-day cycle wants faster decay than a six-month one, where 60 to 90 quiet days can be normal mid-deal while legal reviews the contract.
Both, if you sell to a committee. Contact scores tell a rep who to call; the account score tells you whether a buying group is forming. Five people from one company each sitting in the nurture band is a stronger signal than one person in the sales band, and a contact-only model will never show you that. Roll contact scores up to the account, weight seniority, and let a multi-stakeholder account clear the threshold even when no single contact does.
Rules first, in almost every case. A rules model is readable: a rep can see why a lead scored 80, argue with it, and sometimes be right — which is how the model gets better. Predictive scoring needs a real history of won and lost deals with clean fields before its patterns mean anything, so a team in its first year of inbound gets a confident-looking number built on very little. Start with rules, keep the data tidy while they run, and bring prediction in once there is history to learn from.
Because nothing in the process re-derives them. The business moves upmarket, a new channel brings a different kind of lead, a form field changes and a signal quietly stops firing — and the weights stay where a workshop left them. The symptoms are recognisable: MQL volume looks healthy while MQL-to-opportunity conversion slides, reps start ignoring the score, and high scorers cluster in one source. The fix is the quarterly pass in the steps above, plus a monthly loop where reps name the high-scoring leads that wasted a call.
Quarterly, and as a procedure rather than an opinion. Plot last quarter's won and lost deals against the score they held at handoff; if the two spreads overlap heavily, the model is not separating anything yet. Check each signal's close rate against the current baseline, drop the ones that no longer beat it, and re-derive the rest. Move the threshold only after the weights are right — shifting the cutoff to compensate for bad weights just changes how many wrong leads reach sales.
Yes, but simplify it. At low volume, the value of scoring is not routing speed, it is forcing a shared definition of a good lead between marketing and sales. Use just the firmographic bucket plus the demo-request signal, skip the decay rules, and review every score manually in your weekly pipeline meeting until volume justifies automation.
They complement each other. This rules-based model encodes what your team believes predicts conversion, and the AI score learns patterns from your actual win/loss history. Run both side by side for a quarter: where they disagree, you have either a stale rule worth deleting or a data-quality gap worth fixing.
Ready to use this template?
Deploy B2B Lead Scoring Model in HelloGrowthCRM with one click. Get started free — no credit card · free forever.
Free forever · No credit card required · Set up in minutes
Related resources
Related Templates
Product-Led Growth Scoring
Score users based on product usage to automatically identify Product Qualified Leads (PQLs) for sales outreach.
Inbound Lead Qualification Workflow
End-to-end inbound flow: capture → 0–100 score → BANT → routing → SLA → meeting booking → nurture/recycle, with full automation and field-level steps.