What does lead scoring software actually do?
Lead scoring software puts a number or label on each lead so you know who to contact first. Tools do it in one of three ways, and the way matters more than the brand, because each needs different inputs from you.
| Type | How the score is made | What you must supply | Example |
|---|---|---|---|
| Rules (points) | You decide the points for each property or action | Your own criteria and weights | HubSpot fit and engagement scores |
| Predictive | A model learns from leads you converted in the past | Enough history of won and lost leads | Salesforce Einstein Lead Scoring, 1 to 99 |
| Signal labels | Software reads what a person did or said recently and labels them | A description of who you sell to | Hot, warm or cold with a written reason |
If you haven't written down what a good lead looks like yet, do that first. Our lead scoring framework for teams without a data team takes an afternoon and works in any of these tools.
Which lead scoring tools fit a small team, and what do they cost?
In the first four tools below, lead scoring sits above the entry plan or is sold as an add-on, so the real price is the plan that includes it. Here is what each vendor's own page showed when we checked on 5 November 2026, in US dollars:
| Tool | Scoring type | Plan that includes it | Listed starting price and billing |
|---|---|---|---|
| HubSpot Marketing Hub | Rules; AI-built scores on Enterprise only | Professional or Enterprise | Professional from $800 a month paid annually (the page shows $890 struck through), 3 core seats included, plus a required one-time $3,000 onboarding fee |
| HubSpot Sales Hub | Rules (company and deal scores) | Professional or Enterprise | Professional from $90 per seat a month paid annually (the page shows $100 struck through), plus a required one-time $1,500 onboarding fee |
| Salesforce Sales Cloud | Predictive (Einstein) | Included in Advanced and Max; available for purchase on Core; not in Free, Starter or Pro Suite | Advanced $395 and Max $550 per user a month, billed annually; Core is $195 per user a month, billed annually, plus the scoring add-on, whose price isn't listed |
| ActiveCampaign | Rules | Plus, Pro or Enterprise, through the Pipelines add-on; not on Starter | Plus from $49 a month billed yearly for 1,000 contacts; the add-on price isn't shown on the page |
| Siela | Signal labels | Every account | No plans: $10 of free credit once, then prepaid credit from $50 |
HubSpot's lead scoring help page (last updated 3 November 2026) lists the tool for Marketing Hub and Sales Hub Professional and Enterprise, so the free and Starter plans don't include it. Contact scores need Marketing Hub; Sales Hub scores companies and deals.
For a founder with a few dozen leads a month, the deciding question is whether you need a full marketing platform anyway. If you only want scoring, a $12,600 first year buys a lot of manual prioritising.
How does AI lead scoring work?
Predictive AI scoring looks at leads you converted in the past, finds the patterns they share, and gives new leads a score for how closely they match. Salesforce's Sales Cloud Einstein setup guide (Winter '27, last updated 22 July 2026) says Einstein Lead Scoring scores leads from 1 to 99 based on how well they fit your lead conversion patterns, and that building the model can take up to 48 hours.
The same guide sets a data floor for your own model, for each lead segment you create:
- At least 1,000 leads created in the last 200 days.
- Of those, at least 120 converted to an account and contact.
If you score all leads together and don't meet that floor, Einstein uses a global model built from anonymous data from many Salesforce customers, and once you have enough lead data it builds a model from your data and uses whichever model gives better results. The guide says Einstein Lead Scoring comes with Sales Cloud Einstein and lists it for Enterprise, Performance and Unlimited editions, though its pages differ on whether Performance and Unlimited pay extra. Those edition names don't appear on the current pricing page, which lists Core, Advanced and Max.
HubSpot's help page says Marketing Hub Enterprise can create engagement or fit scores with AI, from a minimum sample of 50 contacts, 25 converted and 25 not; on Professional you set the rules yourself.
A newer approach skips the history entirely: software reads public activity, such as a post asking for a tool recommendation or a new job title, and labels the lead with a reason you can check. Siela's Lead Finder works this way: every lead is scored hot, warm or cold with a one-line reason covering fit, intent, how recent the signal is and how likely they are to reply. Our guide to hot, warm and cold lead signals shows what each label should mean.
Do you have enough data for predictive lead scoring?
Probably not, if you are a small B2B team. Run Salesforce's floor against your own numbers before paying for predictive scoring.
Because the window is rolling, Northwind never qualifies by waiting; only higher volume or a higher conversion rate gets it there. Until then, if it scores all leads together, Einstein would use the global model, which the guide says is built from anonymous data from many Salesforce customers rather than Northwind's own.
That doesn't make the global model useless, but it changes what you are buying. A rules score you wrote yourself, or a signal label with a stated reason, is easier to check at 80 leads a month than a 1 to 99 number trained elsewhere.
How is a lead score calculated?
In a points-based tool, the score is the sum of the points for each rule a lead meets, capped by limits you set. HubSpot's help page describes score groups, each with an optional maximum, inside an overall score limit, and points can be added or subtracted.
Score decay is the part most homemade spreadsheets miss. HubSpot lets engagement points shrink every 1, 3, 6 or 12 months (a month counts as 30 days), in straight-line steps based on the original points.
Decay is what stops last spring's form fills sitting at the top of the list. If your tool has no decay setting, add a "last activity" date and sort by it as well as by score. Whatever the tool, the score is only as good as the fields behind it, which is why CRM data hygiene rules come before any scoring model.
How do you test lead scoring software before you pay?
Back-test it on leads whose outcome you already know. A tool that can't separate last year's winners from losers won't do better with next month's leads.
Export 200 days of leads with outcomes
One row per lead: the fields the tool will score on, the date created, and whether it became a customer, a lost deal or nothing.
Hide the outcomes and score them
Import the file into the trial, or apply your written rules, without the outcome column.
Split into top, middle and bottom thirds
Sort by score and cut the list into three equal groups.
Compare conversion by third
Count customers in each third. If the top third doesn't hold clearly more customers than the bottom third, the score isn't telling you anything yet.
Read five high scores by hand
Ask why each one scored high. If you can't tell, neither can the person who has to call them.
Copy this into the sheet you send to vendors:
If you keep scores next to contacts and deals in one place, the built-in CRM with pipeline and email sequences or your existing HubSpot account both work. If you use HubSpot, our HubSpot lead scoring guide covers setting it up there.
Frequently asked questions
What is lead scoring in AI?
AI lead scoring uses software to rank leads instead of points you assign by hand. Predictive tools learn from your past conversions and give each new lead a score, such as Salesforce Einstein's 1 to 99. Signal-based tools read recent activity, like a public request for a recommendation, and attach a label and a reason. Both still depend on accurate data about who you sell to.
Is lead scoring software worth it for a small business?
Only if the plan that includes scoring is one you would buy anyway. HubSpot Marketing Hub Professional costs from $12,600 in its first year at the price shown for annual payment, with onboarding. With under a hundred leads a month, a written rule set in your CRM, sorted by score and last activity date, often gives the same order of work for far less.
Does HubSpot have free lead scoring?
No. HubSpot's help page, updated 3 November 2026, lists the lead scoring tool for Marketing Hub and Sales Hub on Professional and Enterprise plans. Free and Starter accounts don't include it, so you would track fit and activity in custom properties and sort lists by hand.
What is a good lead score?
There is no universal good number, because every tool and model uses its own scale. Judge a score by what happens next: leads in your top band should convert clearly more often than leads in your bottom band. If they don't, change the criteria rather than the threshold.
Can you do lead scoring in a spreadsheet?
Yes. Add columns for each fit and activity rule, a points formula, and a last activity date, then sort by both. A spreadsheet has no automatic decay, so reduce old engagement points by hand every month or sort by recency. Move to software when updating the sheet takes longer than calling the leads.
Sources
- Salesforce: Set Up Sales Cloud Einstein (Winter '27 guide, last updated 22 July 2026) (checked 4 Nov 2026)
- Salesforce: Sales Cloud pricing (checked 5 Nov 2026)
- HubSpot Knowledge Base: Overview of the lead scoring tool (last updated 3 November 2026) (checked 4 Nov 2026)
- HubSpot: Marketing Hub pricing (checked 5 Nov 2026)
- HubSpot: Sales Hub pricing (checked 5 Nov 2026)
- ActiveCampaign: Pricing (checked 5 Nov 2026)