What this template is for
Use the template to design a first scoring model, to document a model that already runs, or to record a change after calibration. It suits any business that hands leads from marketing to a sales team with limited capacity: business-to-business software, industrial supply, professional services, high-value consumer purchases with a sales conversation. A business where every lead is called the same day has less need for a score than one where sales must choose.
The template is the scoring sheet itself, and three neighbouring pages do different jobs. The lead scoring guide explains signals, weights, decay, negative scoring and calibration in depth, without a sheet to copy. The inbound lead handling playbook runs the process around the score: capture, routing, response service levels, nurture and closed-loop reporting. The lead scoring threshold test designs an experiment to settle where one threshold should sit. The model sheet records the outcome of all three: what the score is, why, and when it last changed.
Keep the sheet in a shared document with a version number and a change log. The platform holds the rules that run; the sheet holds the reasons. When they differ, the sheet is wrong or the platform has been changed without a record, and either needs fixing.
The template
One sheet per scoring model. A business with distinct sales motions, such as new business and expansion, needs one sheet per motion rather than one model that tries to serve both.
| Field | What to write |
|---|---|
| Scope and owner | Which leads the model scores, the system it runs in and who owns changes to it |
| Fit attributes | Firmographic or demographic attributes, the values that count and the points or gate for each |
| Intent behaviours | Actions such as demo requests, quote forms or pricing visits, and the points for each, with caps |
| Decay | How intent points fade with time, and how the rule is implemented |
| Negative scoring | Signals that remove points or disqualify outright: competitors, students, unsubscribes, existing customers |
| Thresholds and stages | The scores at which a lead becomes marketing qualified or is flagged, tied to the agreed stage definitions |
| Routing | Where a lead goes at each threshold, who owns it and the response expected |
| Calibration | How scored leads compared with won and lost outcomes in the last review, what changed and the next review date |
Score fit and intent separately and report both. A single blended number cannot tell a perfect-fit buyer who has done nothing from a poor-fit reader who has done everything, and sales stops trusting it.
How to fill each section
Scope and owner
State which records the model scores (new inbound leads, all contacts, leads from certain sources), which it excludes (existing customers, partners, open opportunities) and the system where the rules run. Name one owner for changes, usually in marketing operations, and name the sales counterpart who agrees them. The lead scoring guide explains why existing customers usually belong in a separate expansion model with different signals.
Fit attributes
List the attributes that describe a good customer: industry, company size, role, region, the problem they have, the equipment or technology they use. For each, write the values that count and the points or gate. Use a few weight levels rather than fine-grained numbers; precision comes from calibration, not from guessing at launch. Out-of-region or out-of-market values are better as gates than as negative points, because a plain condition is easier to explain. The lead scoring guide sets out common fit signals and where their data comes from.
Intent behaviours
List the actions that show buying rather than reading: a quote or demo request, repeated pricing visits, case study or comparison views, live webinar attendance. Give each a weight level and points, and cap the points any one behaviour can earn in a window, so that a colleague refreshing a page or a bot does not push a record over the line. Leave opens out, or nearly so, since mailbox privacy features make them unreliable. Declared high-intent actions, such as a quote request, are better routed immediately than scored; record that rule here and in the routing field.
Decay
Write the rule that makes old behaviour count for less: a rolling window in which only recent intent counts, or a scheduled subtraction after a period of inactivity. Record how the platform implements it and when it was last checked to be running. Fit points do not decay, though they change when the underlying data is refreshed. The lead scoring guide compares the two approaches; the rolling window is simpler to explain and audit.
Negative scoring
List the signals that subtract points and the ones that disqualify outright. Typical groups are disqualifying fit (competitor domains, students and job seekers, personal email addresses in a business-only market), disqualifying behaviour (careers page visits, unsubscribes, bounced addresses) and status (existing customer, open opportunity, recent closed-lost deal). For each, say whether it subtracts or gates. Keep the list short and make each change visible on the record, so a salesperson can see why a score fell.
Thresholds and stages
Write each threshold as a combination of fit and intent, and name the stage it maps to: nurture, engaged, marketing-qualified lead (MQLA lead that marketing judges ready for sales follow-up against agreed criteria.) and the step at which sales accepts the lead and it becomes a Sales-Qualified LeadA lead that sales has reviewed and accepted as a genuine opportunity to pursue.. The stage definitions come from the agreement between the teams, which the MQL, SQL and lead stages guide covers; the model decides who crosses the lines, not what the lines mean. Set first thresholds from where past won leads sat before they were accepted, not from instinct.
Routing
For each threshold, write where the lead goes, how the owner is chosen (territory, segment, rotation), the response time agreed and the return path if sales rejects it, with the reason codes used. The inbound lead handling playbook sets out routing, service levels and nurture for the not-ready. A threshold with no routing rule is a number on a record that nobody acts on.
Calibration
Record the last calibration: the period reviewed, how many leads crossed each threshold, what share were accepted and won, and the same for leads below the line. The measure that matters is the Lead-to-Customer RateShare of leads created in a period that became paying customers. above and below the hand-off line; if the two are similar, the score is not separating. Note what changed as a result and the date of the next review. The lead stages and pipeline guide explains how to build the cohort-based stage conversion report that calibration depends on.
Filled example
Below, the model is completed for an invented business. The points, thresholds, figures and names are illustrative.
Ironvale Racking: inbound lead scoring model v3
| Field | Ironvale's model |
|---|---|
| Scope and owner | New inbound leads from the website, trade show scans and paid social lead forms, UK and Ireland. Excludes existing account contacts (routed to account managers) and installers in the partner programme. Runs in the marketing automation platform, synced to the CRM. Owner: marketing operations manager; agreed with the head of sales |
| Fit attributes | Gate: delivery address in UK or Ireland. Industry: warehousing and logistics, manufacturing, wholesale, retail distribution, 20 points; other commercial, 5. Site size given on form: over 2,000 m², 20; 500–2,000 m², 10; under 500 m², 0. Role: operations, facilities or warehouse manager, director, 15; procurement, 10; other, 0. Business email domain, 5. Fit maximum 60 |
| Intent behaviours | Quote request or site survey request: routed immediately, not scored. Racking load calculator completed, 15. Pricing guide downloaded, 15. Product or case study pages, 3 per page, capped at 15 in 30 days. Installation guide downloaded, 5. Webinar attended live, 10; registered only, 2. Email click, 2, capped at 6. Opens, 0 |
| Decay | Rolling 45-day window: only intent from the last 45 days counts. Checked monthly by the operations manager against a sample of ten records |
| Negative scoring | Gates: competitor domains (list maintained by sales), student or academic addresses, unsubscribed or bounced. Subtract 20: careers page visit in the window. Subtract 10: personal email address with no company name. Status gate: open opportunity or closed-lost in the last 90 days, routed to the deal owner instead |
| Thresholds and stages | Nurture: below fit 30 or intent 15. Engaged: fit 30 or more and intent 15 to 29, receives the racking project nurture sequence. MQL: fit 40 or more and intent 30 or more. Sales accepts or rejects within two working days; accepted leads become SQLs |
| Routing | MQL and direct quote requests: assigned by postcode to one of four area sales managers, task due within four working hours. Rejected MQLs return to nurture with a reason code (too small, no project, competitor, wrong country, duplicate). Engaged leads with fit over 50 get a monthly call list for the inside sales coordinator |
| Calibration | Review of January to June 2026: 640 leads crossed the MQL line; 410 accepted, 118 opportunities, 41 won (lead-to-customer rate 6.4%). Below the line: 2,300 leads, 9 won (0.4%). Seven of the nine had used the load calculator twice or more, so the calculator now scores 15 per use, capped at 30. "Too small" was the most common rejection reason, so the site size points were raised for over 2,000 m². Changes made as v3 on 14 July 2026; next review in January 2027 |
Common mistakes
- One blended number. Fit and intent mixed into a single score cannot be explained or acted on. Keep them separate and show both on the record.
- Points without reasons. Values written only in the platform lose their reasoning when the owner changes. Write the reason beside each value on the sheet.
- Precision at launch. Fine-grained weights before any outcome data are guesses presented as science. Start with a few levels and let calibration refine them.
- No decay rule, or one nobody checks. Old behaviour keeps leads at the top of the queue. Write the rule and check that it runs.
- Disqualifiers as small subtractions. A competitor who has read every page can still cross the line. Make clear disqualifiers gates.
- Thresholds without agreed stages. A line marketing set alone means nothing to sales. Tie each threshold to the stage definition both teams signed.
- Scoring declared buying actions. A quote request waiting for its score to cross a line is a lost sale. Route it immediately.
- Never calibrated. A model judged by whether it looks sensible, not by what closed, drifts until sales ignores it. Compare outcomes above and below the line on a fixed cadence.
The other operational templates, including the experiment documents that test a threshold change, are in the templates library.
