Lead scoring gives every lead a number that rises with signs they are a good fit and ready to talk, so your team calls the most promising people first instead of working the list top to bottom. This article answers how to design a simple points model, build it on the platform with a score field and workflows, and keep the score honest over time with decay and regular calibration. The output is a sorted list each morning that tells each salesperson where to start.
What scoring is for, and what it is not
Scoring is a prioritisation tool. It does not decide whether a lead is qualified; that is a yes or no judgement against your criteria. Scoring orders the leads that remain, so when an owner has twenty open leads and two hours, the twenty are ranked.
It is most useful when you have more leads than time. If your team can contact everyone thoroughly within a day, scoring adds little. If leads wait, it decides which ones wait.
Two kinds of points: fit and behaviour
A score that only counts behaviour rewards curious students and competitors who read every page. A score that only counts fit misses the right buyer who is ready today. Combine both.
Fit points come from facts about the lead: their role, company size, location, the service they asked about. These are added once, when the facts are known.
Behaviour points come from what the lead does: opening a proposal, replying to an email, visiting your pricing page, booking a call, attending a webinar. These are added each time the action happens.
A starting model should have no more than ten scored items. Each should be something you believe genuinely changes the odds of a sale.
Fit: decision-maker role 15 points, company of 20 to 200 staff 10 points, service the firm sells most profitably 10 points, outside the regions served minus 30 points. Behaviour: reply to any email 10 points, pricing page visit 8 points, booking page visit without booking 12 points, webinar attended 10 points. Leads above 40 points would appear at the top of the owner's list. These weights are a hypothetical starting point to calibrate against your own won deals.
Before you start
- A list of recent won and lost deals, to check which traits and actions actually preceded a win.
- Fit fields already on your contact records, such as role and company size.
- Tracking on your website so page visits link to known contacts.
- A threshold you will treat as hot, and an agreement on what owners do when a lead crosses it.
How to set up lead scoring on the platform
- Go to Settings, open Custom Fields, select the Contact object and stay on the Fields tab.
- Click Create field and choose a number Field type for the score.
- Enter Lead score as the Field name, choose a folder and describe the points model in Description.
- Check the Live preview, then click Create custom field.
- In Automation, open or create the workflow for one scored signal and add the trigger for that signal.
- Click Add Action, select Math Operation and name it after the signal it scores.
- In Select Field, choose the Lead score custom field.
- Choose Addition for a positive signal or Subtraction for a negative one, then enter the points as the operation value.
- Set the Update Field to Lead score so each result is written back to the contact.
- Click Save, then Publish, and repeat for every signal in your model.
To give owners their sorted list, save a contact view of open leads ordered by Lead score, and add a notification to the owner when a lead crosses your hot threshold.
Before anyone relies on the list, run a calibration check. Score your recent won and lost deals by hand against the model. If leads you won would have scored low, the model is weighting the wrong things. Adjust before the team starts trusting the list.
Decay keeps the score truthful
Without decay, a lead who was very active a year ago still sits near the top today. Build decay as one more workflow: a wait or timed check for inactivity, followed by a Math Operation that subtracts points, so the list reflects current interest. A common approach is to subtract a few points each fortnight of inactivity, down to the lead's fit score but no lower. Fit does not decay; a decision-maker at the right size of company stays one.
What it replaces and the weekly cost
Scoring replaces the salesperson's instinct for who is hot, which is good for the leads they remember and blind to the rest. It replaces the manager asking "who's ready?" in a meeting. Once built, it runs on its own. Allow an hour a month to compare high-scoring leads against outcomes and adjust points.
The mistake most people make
They build a complicated model on day one, with dozens of items and fine-grained points nobody can explain. Salespeople do not trust a number they cannot reason about, so they ignore it. Start with a handful of items you can justify in a sentence each, and add more only when the outcomes show a gap.
Scoring email opens. Privacy features in mail apps register opens that never happened, so open-based points inflate scores at random. Score replies and clicks instead.
When not to use lead scoring
If you receive few leads and each gets full attention, scoring adds maintenance without changing anyone's day. It also misleads in very long, relationship-based sales where most of the signal is in conversations the platform cannot see. In those cases a short qualification checklist, described in qualifying leads, serves you better.
Troubleshooting
Scores climb endlessly. A behaviour workflow fires every time a page loads. Limit it to once per day or once per contact with a condition.
Every lead has the same score. Fit fields are empty, so only behaviour counts. Fill fit fields through the form or an AI agent's questions.
Hot alerts arrive for existing customers. The model does not exclude clients. Add a condition that skips contacts tagged as customers.
Owners ignore the sorted view. It is not their starting screen. Make it their default list and review it together for two weeks.
Frequently asked questions
Is AI lead scoring better than points?
It can find patterns you would miss, but it needs a meaningful history of won and lost deals to learn from. For most mid-sized firms, a clear points model gets you most of the value and is easier to trust.
Should scoring be visible to the lead?
No. The score is an internal priority signal. Treat it like any internal note.
How often should we recalibrate?
Monthly for the first quarter, then quarterly. Recalibrate sooner if you launch a new service or marketing channel, because the leads change.
Does scoring apply to leads already in nurture?
Yes, and that is one of its best uses. A nurtured lead whose score jumps is showing intent again. The sequence design is covered in setting up lead nurturing, and returning leads should be routed back to their owner as set out in lead management.