How to Detect Fake Leads Using Email Data
Fake and low-quality leads are best caught by checking the submitted email with the b2b_lead (or crm) context specifically, not the default generic context — role-based addresses and free-provider addresses are weighted far more heavily there, because a named individual on a company domain is what most B2B forms actually need, and the default weighting doesn't reflect that.
Why lead quality isn't the same problem as signup screening
A consumer signup form mostly cares whether an address is real and not disposable — a free Gmail address is completely normal there. A B2B lead form has a different, narrower definition of "good": ideally a named individual, reachable on a real company domain, not a shared inbox. sales@bigcompany.com might be perfectly deliverable and completely legitimate, and still be a low-quality lead for a sales team trying to reach a specific decision-maker.
Why the context parameter is the actual fix here
Rather than writing separate lead-scoring logic, the same risk engine handles this correctly just by passing the right context. In the b2b_lead context specifically, role-based addresses are weighted 2.5× and free-provider addresses 2× relative to the generic context — a role-based free-provider combination that would barely register on a consumer signup form becomes a meaningfully higher score on a lead form, which is the correct behavior for that use case. The crm context applies a similar but slightly gentler adjustment, for cases where you're importing contacts into a CRM rather than screening a live lead-capture form.
The same address, two contexts
Take info@company.com — role-based, on a legitimate, established company domain, with clean SPF/DMARC:
| Context | role_based_address contribution | Likely outcome |
|---|---|---|
| generic | 10 points (base weight) | Low risk — barely registers |
| b2b_lead | 25 points (10 × 2.5) | On its own, likely enough to land in review — worth a second look before routing to sales |
Nothing about the address changed — only what "acceptable" means for the form it was submitted to.
Signals worth combining, beyond the score itself
- domain.type — business vs consumer vs disposable — a lead on a business domain is inherently a stronger signal than one on a free consumer provider, independent of the role-based check
- domain.infrastructure — whether the domain's mail is on known business infrastructure (Google Workspace, Microsoft 365-style setups) vs unknown/self-hosted, which correlates with how established the company is
- local_part_analysis.name_like — whether the local part reads as an actual person's name rather than a generic or generated string, useful context for a sales team deciding how much to trust a lead before outreach
What not to do
- Don't hard-reject every role-based or free-provider lead — many genuinely good leads come through general inboxes or personal accounts, especially from smaller companies without individual work email conventions
- Don't use the generic context for a lead form and then complain the scoring "doesn't catch" role-based addresses — the context parameter exists specifically to fix this without custom logic
- Don't treat lead quality scoring as a substitute for actual lead qualification (budget, authority, need, timeline) — it filters out the address-level noise, not the sales-level fit
FAQ
No — many legitimate contacts, especially from smaller companies or in early research stages, use personal addresses. It's a weighted signal, not a disqualifier.
Both weight role-based and free-provider signals up relative to generic, with b2b_lead applying a slightly stronger adjustment — suited to live lead-capture forms versus general CRM contact import.
No — it filters address-level quality (is this a reachable, individually-attributable contact) before a lead ever reaches a rep. Actual qualification (fit, intent, budget) is a separate process.
Usually not outright — route them to review rather than hard-blocking, since a role-based address on an otherwise strong company domain can still be a legitimate entry point into a real account.