Email, Marketing & Creator Tools

Lead Scoring Calculator

Design a lead scoring model: fit and engagement attributes with points, decay and negative scoring, MQL thresholds, then score a CSV of leads and see the distribution before you configure your CRM.

Scored leads
Score distribution
Rules JSON and CSV

Runs in your browser

Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics.

Examples:

1 Scoring rules

One rule per line

fit: job_title contains director => 15
engagement: pricing_views per 1 => 5 max 20

fit: or engagement:, then a field, a test and => points. Tests: = value, != value, contains text, in (a, b), >= 100, <= 100, between 50 and 500, exists, empty, per 1 (points for each unit).

Optional endings: max 20 caps a rule; half-life 30 days from last_visit halves the points every 30 days since that date. Negative points are allowed. # starts a comment.

MQL threshold (total score)

Minimum fit score for MQL (optional)

Score as of

Distribution bucket width

2 Leads to score

Leads as CSV with a header row

name,job_title,pricing_views
Sample lead,Director of Sales,3

Column names must match the fields in your rules (case does not matter). Dates as YYYY-MM-DD. Up to 50,000 leads.

Drop a CRM export here or choose a file (CSV, up to 10 MB)

3 Scores

Write your rules, paste leads and press Score leads.

What the Lead Scoring Calculator does

This lead scoring calculator lets you design a points-based scoring model and test it on real leads before you configure it in your CRM or marketing automation platform. You write fit rules (who the lead is) and engagement rules (what they did), with negative points, caps and time decay, set an MQL threshold, then score a CSV export and see the distribution, the MQL count and exactly why each lead scored what it did.

The score is a transparent sum of your rules. Nothing is learned from data and nothing predicts who will buy; the value of the tool is that you can see the consequences of a model - how many MQLs it would create, and which leads it would miss - before sales has to live with it.

How to use it

Write one rule per line, starting with fit: or engagement:, then a field, a test and the points, such as fit: company_size between 50 and 1000 => 20.
Use per for counts (engagement: pricing_views per 1 => 5 max 20), negative points for disqualifiers (fit: email contains gmail.com => -15) and half-life 30 days from demo_date for activity that should fade over time.
Set the MQL threshold, and optionally a minimum fit score so that a very active student or competitor cannot become an MQL on engagement alone.
Paste or drop a CSV export of leads whose column names match your rule fields, choose the as-of date and press Score leads.
Read the distribution and the MQL count, open Explain on any lead to see every rule that fired, then download the scored CSV or the rules as JSON to recreate in your platform.

Reading the results

Fit measures how closely a lead matches the customers you want; engagement measures interest shown recently. Keeping them separate stops a perfect-fit lead who never engaged, or an eager lead who will never buy, from looking the same.

The distribution shows how the threshold cuts your database. If half your leads are MQLs, the threshold is probably too low for sales to follow up; if almost none are, it may be too high or the rules too strict.

Half-life decay multiplies a rule's points by 0.5 for every half-life that has passed since the activity date, so a demo request is worth full points today, half after 30 days and a quarter after 60.

Worked example: scoring one B2B lead

Priya is a Director of Marketing at a 240-person company in the UK, viewed the pricing page 4 times and requested a demo on 12 September. The model is scored as of 19 September.

Fit: the job title contains director (+15), the company size is between 50 and 1,000 (+20) and the country is in the target list (+10), so fit is 45.

Engagement: 4 pricing views at 5 points each give 20, within the cap of 20. The demo request is worth 40 with a 30-day half-life; it is 7 days old, so it counts 40 x 0.5^(7/30) = 40 x 0.8507 = 34.0. Engagement is 54.0 and the total 99.0, above the MQL threshold of 60 and the minimum fit of 25, so Priya is an MQL.

Formulas and scoring rules

Total score: total = sum of fit rule points + sum of engagement rule points
Counting rule: points = floor(value / unit) x points per unit, then capped at max
Time decay: counted = points x 0.5 ^ (age in days / half-life)Age is measured from the rule's date field to the as-of date; a missing or invalid date counts as zero points.
MQL: MQL when total >= threshold and (no minimum fit, or fit >= minimum fit)

Building a model sales will trust

Start from closed deals, not from opinions: list the job titles, company sizes and actions that your last 20 or 30 customers had in common, and give those the most points. Then run the model on last quarter's leads and ask sales which of the top-scored leads were genuinely worth a call. Adjust until the answer is mostly yes.

Negative scoring is as important as positive. Personal email domains, students, competitors, existing customers and people who unsubscribed should lose points, or be excluded outright, so they do not crowd the MQL queue.

Limitations: what the result does not prove

It scores with the rules you write; it does not learn which attributes predict conversion and cannot tell you whether your model is right.
CRMs implement decay, caps and thresholds differently. Recreate the model in your platform and check a few leads by hand against the scores here.
Text matching is literal after lower-casing: director does not match Dir.. Clean or map job titles first if your data is messy.
Lead data is handled only in your browser. Even so, export only the columns the rules need.

Privacy: where your data goes

Everything you paste, type or drop is processed in this browser tab. It is not uploaded, logged, stored or sent to analytics. Session recording and tag-manager scripts are switched off on this page.

Standards and sources

Frequently asked questions

What is the difference between fit and engagement scoring?

Fit, sometimes called explicit or demographic scoring, rates who the lead is: role, company size, industry, country. Engagement, or behavioural scoring, rates what they did: page views, downloads, webinar attendance, demo requests. Many teams require a minimum on both before passing a lead to sales.

What MQL threshold should I use?

There is no standard number, because the scale depends entirely on your rules. Pick the threshold by looking at the distribution: choose a cut that sends sales a volume they can follow up within a day or two, then check that the leads above it are ones sales actually wants.

How does lead score decay work?

Decay reduces the points for an activity as it ages, so last week's demo request outweighs one from last spring. This tool uses a half-life: after one half-life the activity is worth half its points, after two a quarter. Some platforms instead subtract fixed points each month; the effect is similar.

Should lead scores ever be negative?

Yes. Negative points are the simplest way to push out leads you should not call: competitors, students, personal email addresses, job seekers and people who unsubscribed. A lead whose total drops below zero is marked for review rather than silently kept in the queue.

Can I import these rules into HubSpot, Salesforce or Marketo?

Not directly - each platform has its own rule editor. Download the rules as JSON or copy them as text and recreate each line in your platform's scoring settings. Then score a handful of the same leads in both places to confirm they agree.

How many rules should a lead scoring model have?

Fewer than you might think. Ten to twenty rules covering your clearest fit signals, your strongest intent actions and your main disqualifiers are easier to explain and maintain than a hundred small ones. Each rule here shows how often it fired, so you can remove rules that never matter.

Last reviewed 19 September 2026 by the A2Z.Tools team against the sources listed above.