# Lead scoring: a practical worksheet for fit, intent and follow-up

Build a simple lead-scoring model without confusing activity with buying intent. Includes a fictional scoring example, review checklist and handoff rules.

Canonical page: https://cpcinsider.com/insights/lead-scoring-worksheet/
Author: Vladlens Kecko
Published: 2026-11-04
Updated: 2026-11-04


**Lead scoring assigns values to prospect information or actions to help a team prioritise follow-up.** It should make the next action clearer, not turn an arbitrary total into a claim about purchase probability. Start with a small, explainable worksheet before automating decisions.

[Salesforce’s lead-scoring guide](https://www.salesforce.com/blog/sales/lead-scoring/) describes ranking prospects using their characteristics and behaviour. Our worksheet below applies that concept to an ordinary B2B enquiry queue. It complements the wider [lead-generation system](https://cpcinsider.com/insights/b2b-lead-generation-system/); it does not replace sales qualification.

## 1. Define the decision the score supports

Choose one offer and one team. Write the output as a decision: “Which eligible enquiries need a sales conversation now, and which need more information?” Agree an owner, a backup and a response expectation the team can actually meet.

Handle explicit requests for help promptly through the normal enquiry process. Do not make someone wait for a score to accumulate after they have asked for a quote. Existing customers with a service problem belong in the appropriate support process, not a prospect-ranking contest.

## 2. Keep customer fit and engagement separate

Fit concerns whether the company can use the service you deliver. Engagement concerns the actions you have reliably observed. [HubSpot’s scoring documentation](https://knowledge.hubspot.com/scoring/understand-the-lead-scoring-tool) distinguishes fit, engagement and combined scores. This is a useful distinction even if your first version is a spreadsheet rather than HubSpot.

Use relevant business facts such as service need and delivery coverage. An unknown field means “needs checking”, not “poor fit”. Repeated page views can have many explanations; they are not equivalent to an agreed project. Keep essential eligibility rules outside a total that irrelevant clicks could overwhelm.

## 3. Try this fictional scoring model

These weights and thresholds are invented for a maintenance service. They are not market benchmarks, a validated prediction model or a recommended default for every company.

| Dimension | Confirmed signal | Example points |
| --- | --- | --- |
| Fit | Requested service is offered | 20 |
| Fit | Site is inside the delivery area | 10 |
| Engagement | Requested a scoped quotation | 20 |
| Engagement | Attended an agreed discovery meeting | 10 |

A fictional enquiry with the first three signals scores **30 fit + 20 engagement = 50 total**. That does not mean a 50% chance of buying. Keep the two subtotals visible. A second prospect with many low-value interactions must not outrank a genuinely suitable buyer just because activity was counted repeatedly.

Count each listed signal once in this prototype. Record when it happened. Choose how old evidence expires before using the worksheet; do not let a years-old interaction imply current readiness. Where information conflicts, ask the owner to review it.

## 4. Turn the model into routing rules

| Situation | Practical next action |
| --- | --- |
| Suitable company, explicit current request | Assign a sales owner and confirm the next step |
| Suitable company, early research only | Offer the appropriate information; record a review condition |
| Important fit field missing | Verify it before classifying the enquiry |
| Confirmed request outside service scope | Explain the limitation; record the reason |
| Existing customer needs assistance | Route to the existing customer-service process |

The table is a suggested workflow, not a CRM setting we have enabled. Record the reason for each route, so sales can challenge a misleading score. A numerical threshold cannot repair missing [CRM ownership and measurement](https://cpcinsider.com/insights/crm-conversion-tracking/).

## 5. Test the logic before switching on automation

Review a small, relevant set of historical records with known outcomes. Reconstruct only information available at the original decision time: using the eventual “won” status as an input would make the test misleading. Compare suitable enquiries the model prioritises with suitable enquiries it misses.

Document missing data, duplicates and exceptions. Inspect whether point totals reflect useful evidence or merely repeat activity. If using HubSpot, its documentation provides record testing and score-distribution preview; feature access depends on the account. This guide does not require a new paid tool.

## 6. Review quality, workload and missed opportunities

Use one review sheet: eligible enquiries, priority decisions, accepted sales conversations, rejection reasons, response workload and subsequent outcomes over a suitable sales window. Preserve the rule version. A change in conversion after scoring is introduced can also reflect demand, offer or sales changes; it is not automatic proof of causal lift.

For campaign-level diagnosis, use the [Meta lead-quality checklist](https://cpcinsider.com/insights/meta-ads-lead-quality/). Keep the model simple enough that the team can explain a decision to a colleague.

## Make the handoff easier to operate

The output is one documented model, one routing agreement and one review date. Explore [B2B lead-generation support](https://cpcinsider.com/services/b2b-lead-generation/) or [request a free 30-minute founder-led growth review](https://cpcinsider.com/#audit). You receive a short written plan with three priorities, with no obligation to hire CPCInsider.