B2B lead scoring: a model for sales and a signal for your ads

What is a B2B lead scoring model?

The classic model adds points for who the lead is and for what they have done: visited the pricing page, opened emails, downloaded a guide. The trouble in B2B is that engagement is easy to produce and weakly tied to buying. A procurement manager with a live RFQ may visit one page and fill in one form. A student researching a thesis may read ten articles and open every email.

So this guide weights requirement and intent above engagement, and treats the score as a prediction that has to be checked against what later happened. What counts as a qualified lead in the first place is covered in B2B lead quality; this page is about turning that definition into a score and a signal.

Why does one score fail both sales and the ad platforms?

Send a 0–100 score that sales re-weights every month to an ad platform and you get two problems. The meaning of “72” changes under the bidding system while it is still learning, and the score often depends on inputs that arrive days after the click. Give sales only four coarse tiers, on the other hand, and they lose the detail that helps them decide who to call first.

The answer is to keep both, and connect them deliberately. We call it the Score-to-Signal Map.

Which signals should a B2B lead score use?

The best signals are ones a lead states or sales confirms, because they don’t depend on how often you advertised to that person. They differ by segment:

Signals to leave out or discount

• Email opens. Apple’s Mail Privacy Protection downloads remote content “in the background by default — regardless of whether you engage with the email” (Apple). An “open” can be recorded without anyone reading the email.
• Page views your ads created. A pricing-page visit from a retargeting ad tells you the ad worked, not that the visitor is buying. Scoring on it rewards the campaign for its own traffic.
• Hard penalties for personal email addresses in India. Owners of smaller firms and many industrial buyers use Gmail or Yahoo. A GSTIN or a company name is a better fit test than the email domain.
• Large negative points for invalid leads. A student with −50 points is still in the pool. Give invalid leads their own status so they are excluded, not just ranked low.

If a signal isn’t captured today, the form is the place to add it. See form and capture optimisation for how to ask without losing good leads, and LinkedIn Lead Gen Forms lead quality for custom questions on pre-filled forms. On Google Search lead forms, qualifying responses let you mark particular answers as qualifying, and Google tags those submissions as qualified leads automatically (Google Ads Help).

How do you set tiers and cut-offs?

Here is a first model for an Indian manufacturer of industrial valves selling through Google Search and LinkedIn. The maximum score is 100.

An EPC contractor asking for 40 DN150 valves against a minimum order of 25, delivered to a served state within 60 days, with a GSTIN, scores 30 + 20 + 20 + 10 + 10 + 10 = 100: Priority. A distributor with a matching specification but only 10 units, no timeline and a Gmail address scores 30 + 0 + 10 + 10 + 0 + 0 = 50: Qualified. Response times per tier belong in the sales handoff guide; the point here is that each tier is a status sales can record.

How do you turn tiers into values for value-based bidding?

Google describes value-based bidding as Smart Bidding that optimises to the value conversions bring, using Maximise conversion value with or without a target ROAS, and says values can be proxies such as a lead score rather than revenue (Google Ads Help). Its best-practice page adds that particular lead-form answers may indicate higher-quality leads and can carry different values (Google Ads Help).

Continuing the valve manufacturer, with rates taken from past RFQ cohorts and an average purchase order of ₹8,00,000:

Only Priority and Qualified are sent, as one qualified-lead conversion action, with values of ₹60,000 and ₹16,000 (or 25 and 6.7). Three cautions:

• These are expected values, not revenue. A conversion value of ₹1,52,000 from two Priority and two Qualified leads in a week is a forecast. Keep actual pipeline and orders in a separate report.
• Watch the spread. A 25-to-1 ratio tells bidding to chase Priority leads hard. If Priority leads are rare, bidding may overpay for the few auctions it associates with them. An index with a narrower spread is a reasonable compromise while volume is low.
• Conversion value rules don’t replace this. Google’s value rules adjust values by location, device and audience at auction time (Google Ads Help). They don’t know what a lead told your sales team, so tier values have to come from your own data.

If you aren’t sure values are warranted yet, send the qualified stage without values first. The trade-offs are in which lead stage to optimise for, and Smart Bidding for lead generation covers the bid strategy side.

How do you send score tiers to Google, Meta and LinkedIn?

Meta’s 1–40% range is a useful test of tier design. If Qualified and Priority together cover 60% of leads, the tier is too loose to teach Meta much; if they cover 0.5%, it is too rare. Under 200 leads a month, don’t send the score to Meta at all: use qualifying questions on the Instant Form and review quality by hand (see Instant Forms lead quality).

Upload mechanics, identifiers and stage naming are in CRM lead stages for ad platforms. Zephra sends CRM lead stages to Google (Data Manager API), Meta (Conversions API) and LinkedIn (Conversions API) where configured, and works from CRM data or a spreadsheet, so no CRM is needed to start. See lead quality optimisation.

How do you know the lead scoring model works?

Build a validation tab with one row per tier and a cohort window, such as leads created January to March and checked in June:

The tiers separate well: Priority converts at about 2.6 times the Qualified rate and about 11 times the Low rate. That is enough to keep the model. It isn’t enough to fine-tune individual weights, because each tier has fewer than 30 quotes. We use 30 outcomes per tier as a rule of thumb, not a statistical law; below it, two or three quotes either way move a rate by several points.

Decide what to do with the result using the same four decisions Zephra applies to campaigns:

Leads with no status yet must stay out of the denominator. Counting “unknown” as “didn’t convert” makes every tier look worse and the newest cohort look worst. The lead quality guide covers small counts and cohort maturity in more depth.

Rule-based or predictive lead scoring: which do you need?

Buying committees change the unit. When three people from one company enquire, score the account as well as each contact, and don’t send three Priority conversions for one opportunity. For account lists on LinkedIn, see ABM targeting.

Market notes

• India. Don’t penalise Gmail or Yahoo addresses; use GSTIN or company name as the fit test. WhatsApp-only enquiries need a tier too, or they vanish from validation. Marketplace enquiries can be scored with the same sheet for comparison (see marketplaces vs paid ads).
• UAE and GCC. A trade licence or free-zone entity can be a fit signal. Agents and intermediaries need an explicit rule: decide whether an agent enquiring for an end buyer is Qualified or Low. Score by country, because costs and buyer mix differ across the GCC (see GCC B2B paid ads).
• US. MQL thresholds are usually score-based and SQL acceptance is a sales decision. For ad platforms, a sales-accepted stage is normally a more trustworthy signal than a marketing score alone.

Lead scoring template and what to measure next

No CRM is needed. Build a sheet with four tabs:

Then track, by campaign and channel:

• Tier mix: share of leads in Priority and Qualified.
• Tier conversion rate: opportunities or quotes ÷ leads, per tier, on matured cohorts.
• Cost per qualified lead: spend ÷ Priority plus Qualified leads. Use the cost per qualified lead calculator.
• Share of spend producing Priority leads: which campaigns earn their budget.
• Unscored or unknown rate: leads with no tier or status after your agreed response time.

For the wider system these measures sit in, see the B2B lead generation playbook. Before uploading lead data to ad platforms, check lead generation compliance.