TL;DR — The lead scoring B2B is a system that assigns a score to each lead by combining fit (how much it is in target) and engagement (how much it interacted). It only works if it is calibrated on customers you have already closed, not on a template found online — and it must be recalibrated in the first 60-90 days, not left fixed. Under the 40-50 lead per month, however, a formal model is often more setup work than value: a manual A/B/C tier managed by a senior SDR does the same job, before and without extra software.
When leads begin to arrive from multiple channels together — form on the site, outbound campaigns, events, referral — the question that almost every B2B SME is always the same: on what is worth calling first? This is where the lead scoring B2B: a system that assigns a numerical score to each contact based on two things — who it is, and what it did — to separate who deserves a call today from those who should be let mature.
The problem is that most online guides treat lead scoring as if it were a universal exercise, equal for an e-commerce B2C and for a manufacturing SMEs that closes 8 customers a year. It's not. In this guide we see how to build a model of lead generation B2B that really works, how much they cost the tools in 2026, and — point that almost nobody treats — when for an SME it does not yet serve, and a manual tier does the same job better.
What is the lead scoring B2B: the two dimensions that make it
The lead scoring combines two types of scores, and confuse them is the most common mistake:
- Fit scoring — how much contact looks like your ideal customer: role in company, size, industry, technological stack. It is a static photograph, it rarely changes.
- Engagement scoring — how much contact interacted with you: open emails, pages visited, answers to a touchpoint, participation in a webinar. It's dynamic, it changes every week.
A lead with high fit but zero engagement is a great target for a campaign outbound Targeted, not a lead ready for commercial. A lead with high engagement but low fit — downloaded three guides but working in a company too small for your average ticket — is the classic case in which the rough score mind, if you do not correctly weigh the two sizes together.
Because many Italian SMEs do not need it yet
This is the point that most of the content on lead scoring jumps because it does not sell software: if you generate 20-30 leads per month, a model of formal scoring — weights, thresholds, a motor that autocalculates them — it is more setup time and maintenance than it saves you. With those volumes, an experienced commercial reads a lead and understands in 30 seconds if it is worth a call; the numerical score tells him nothing he doesn't know already.
The lead scoring pays off when one of these two things happens: the leads come from more sources together (site, outbound, events, referral) and no one has the time to read them one by one, or the volume exceeds what a person can sort in mind — indicatively, over 40-50 requests per month. Below that threshold, honest shortcut is a manual tier: A for those who are clearly in target and showed interest, B for those who are in target but cold, C for the rest. Less elegant than a point model, but calibrated in an afternoon instead of in a quarter.
How to build a lead scoring model in practice
The fastest way to build a useless model is to copy one from a blog and apply it as it is. It only works if you leave. Yours customers, not from a generic example:
Step 1 — Analyze closed customers in the last 12 months
Before you assign one point, look who really bought it. What roles do decision makers play? What size did companies have? What actions had they done before becoming customers (visited the price page, responded to an email, participated in a demo)? This is the dataset that defines weights — not the intuition, not a template.
Step 2 — Assign weights to fit and engagement separately
An example of 100-point structure, designed for a B2B sales cycle with a medium-high ticket (adapted to your data, not copied):
| Criteria | Size | Weight |
|---|---|---|
| Company in ICP dimensional range | Fit | 25 p.m. |
| Role in buying committee (decider or influencer) | Fit | 20 p.m. |
| Target sector | Fit | 10 p.m. |
| He responded to a touchpoint (email, LinkedIn, phone) | Engagement | 20 p.m. |
| He visited the prices/services page | Engagement | 15 p.m. |
| He participated in a call or demo | Engagement | 10 p.m. |
Step 3 — Fix the thresholds, then test them seriously
A typical structure: under 40 points the lead remains in nurturing, between 40 and 70 becomes MQL (marketing qualified lead, to work but not urgent), over 70 becomes SQL (sales qualified lead, goes to commercial by the day). But the exact number counts less than the next step: in the first 60-90 days, check how many SQL become customers. If the rate is too low, the thresholds are too permissive; if the commercial complains that it receives few leads, they are too narrow.
Step 4 — Ricalibra, do not leave the fixed model
The most popular lead scoring model in Italian SMEs is the one set once and never touched again. It is also the reason why, after a year, almost no one trusts the score: the market changes, the ICP gets refined, and a static model ages faster than it looks.
Manual, in CRM or with AI: three ways to make comparison scoring
We, in practice, work like this: we build segmented lists in A/B/C tier with Clay and automatic enrichment, and each lead that answers however passes from a BANT qualification made by a senior SDR before arriving to the customer — a hybrid between automatic scoring upstream and human judgment downstream, not a blind score that decides alone who deserves a phone call.
| Approach | How it works | When it makes sense |
|---|---|---|
| Manual (tier A/B/C) | A commercial or SDR rankings in judgment, often on spreadsheet or directly in CRM | Low volumes (<40-50 lead/month), ICP simple to recognize to the eye |
| Native CRM Scoring | Automatic score on rules (visits, emails, fields) in Pipedrive or HubSpot | CRM already in use, average volumes, it only serves to sort without changing stack |
| enrichment layer + AI | It enriches the lead with external data and segments before it enters CRM, often via AI automation | High volumes, multiple sources to unify (form + outbound + events) |
How much lead scoring costs in 2026 (real numbers)
If your CRM already includes it, the additional cost is zero — it is just a question of configuration. If you have to step up to unlock it, the numbers change a lot from one supplier to another:
| Instrument | Plan that includes scoring | Indicative cost |
|---|---|---|
| Pipedrive | Growth (AI Engagement Scoring via Pulse) | ~39 €/user/month, annual billing |
| HubSpot Sales Hub | Professional (contact/deal scoring) | from $ 90/seat/month, annual billing |
| Manual calculation sheet | — | Free, but requires recurring management time |
Source: HubSpot official price page (hubspot.com/pricing/sales) and public pricing guides on Pipedrive Growth, consulted in September 2026 — prices vary by number of users and must be verified on the list at the time of purchase.
The account that really counts is not the monthly fee, but the time that a bad model loses to the commercial — both chasing leads that will never close, and ignoring good leads because the score has classified them badly. A simple model, calibrated on your real customers, almost always beats a complex copied from another sector.
Lead Scoring and GDPR: what you need to know
Point that almost no lead scoring guide treats precisely: assigning a score to a contact based on his behavior is, to all effects, a treatment profiling pursuant to EU Regulation 2016/679 (GDPR). You do not need a separate consent to do so on data that you have already collected legitimately — for example within the limits of the legitimate interest that regulate most of the cold B2B email in Italy — but if the system is systematic and operated on a large scale, theArticle 35 of the Regulation imposes an impact assessment on data protection (DPIA). It is a passage that concerns above all those who scale the lead scoring on thousands of contacts with an automatic motor, not the PMI that holds a manual tier on a hundred leads — but it is the difference between a compliant system and one that no one has ever controlled.
The most common mistakes when building a lead scoring model
- Copy a model found online without calibrating it on your customers. The "standard" weights you spin for the network rarely reflect your real sales cycle.
- Give more weight to the engagement than to the fit. A contact that clicks everything but has never been in target remains an off target contact, however active.
- Never recalibrate the thresholds after the launch. A model fixed once and never revised loses accuracy faster than it looks.
- Keep the scoring isolated from CRM or outbound sequence. If the score does not automatically update where the sales works, no one really looks at it — exactly the same problem we face when we connect CRM and operating tools for customers.
From where to start: the framework in 3 moves
- Start with closed customers, not a template. Look at the last 12 months before assigning one point.
- Choose a few thresholds and heads 60-90 days before trusting you. The exact number counts less than the verification cycle after launch.
- Connect it to where your team already works. CRM or outbound sequence — an isolated scoring is a scoring that no one will use.
It is the same principle with which we build systems lead generation B2B for our customers: segmentation per tier before the lead enters sequence, and a human qualification — not only a score — before it arrives at the commercial. If the bottleneck of your funnel is not generating lead but sorting them in time, it is also worth reading our guide on because so many contacts produce few appointments.
FAQ: lead scoring B2B
What is lead scoring B2B?
It is a system that assigns a numerical score to each lead by combining two dimensions: the fit (when the contact corresponds to your ideal customer — role, company size, industry) and the engagement (how much it interacted with you — open emails, visited pages, answers). The score is to decide who to contact first and who to leave in nurturing.
How do you build a lead scoring model?
Starting from customers closed in the last 12 months, not from a template found online: they isolate the characteristics and behaviors that real customers had in common, they allocate a weight to each, and fix thresholds (MQL, SQL) that are tested and recalibrated in the first 60-90 days, do not leave firm forever.
Is the lead scoring manual or automatic more suitable for an SME?
Depends on the volume. Under 40-50 leads per month, a manual tier (A/B/C) managed by a commercial or senior SDR often works better than an automatic CRM model: it is faster to calibrate and does not require additional software. The automatic model repays when leads come from multiple sources together and no one has time to read them one by one.
Does the lead scoring require GDPR consent?
Lead scoring is a profiling treatment under the GDPR: no separate consent is required to assign a score to data already collected legitimately, but if the system is systematic and on a large scale Article 35 of the Regulation requires an impact assessment (DPIA) — a passage that most lead scoring guides do not mention either.
How much does a lead scoring tool cost?
If your CRM already includes it, the additional cost is zero: Pipedrive offers it from the Growth plan (about 39 euro/user/month with annual billing), HubSpot Sales Hub unlocks it from the Professional plan (from 90 dollars/seat/month with annual billing). The manual alternative on spreadsheet is free, but costs management time that is to be put into account.
Lead scoring and lead qualification (BANT) are the same?
No. The lead scoring is an automatic filter that orders the leads for priority before someone looks at them. The BANT qualification (Budget, Authority, Need, Timeline) is the final verification, made by a person, before passing the lead to the commercial. The first reduces noise, the second confirms that it is really worth investing time.
Related Insight: Lead generation B2B: what is it and how it works for the complete picture of the funnel, and integrate CRM and billing to connect the scoring to where your team already works.