Data Methodology

Pristine aggregates contact records across three primary enrichment providers: Wiza, Explorium, and a third-party database. Each provider maintains its own crawled and contributed dataset. Our figure of 700M+ represents the deduplicated union of all three sources after deduplication by email domain and LinkedIn URL.

This number is audited quarterly. Records older than 18 months without a verification event are flagged and excluded from active search results, keeping the searchable database conservative rather than inflated.

Deliverability is measured as the percentage of exported contacts that pass real-time SMTP verification at the moment of export. An email address that fails SMTP handshake is filtered before it reaches the user. You never see a contact we cannot verify.

The 90% figure is a floor, not an average. It reflects our minimum guarantee across all export cohorts. In practice, most campaigns see 92–96% deliverability depending on the ICP searched. We measure this by tracking hard bounces on outbound campaigns sent through Smartlead and Instantly integrations, aggregated across all customer accounts monthly.

Comparison context: industry bounce rate benchmarks for cold email typically run 3–8% on clean lists and 10–20%+ on unverified database exports (source: Smartlead industry report, 2024; Instantly deliverability benchmarks, 2024).

When a user exports a contact, Pristine performs a live SMTP handshake against the mail server for that email domain. This is a three-step process: (1) DNS MX record lookup to identify the mail server, (2) SMTP EHLO/HELO handshake to confirm the server is reachable, (3) RCPT TO command to verify the specific mailbox exists, without sending an email.

This process takes 300–800ms per contact and runs in parallel across batches. It is performed at export time, not at ingestion time, which means the verification reflects the current state of the mailbox rather than a cached result from weeks or months prior.

Limitations: SMTP verification cannot detect catch-all domains (domains that accept mail for any address). We flag catch-all domains in the UI so users can decide how to treat them. Catch-all domains represent approximately 8–12% of the B2B contact universe.

When a user searches, Pristine queries Wiza, Explorium, and our third-party database simultaneously. Each provider returns its best available data for each contact. Our resolution layer then selects the highest-confidence value per field, not per provider.

For example: if one provider has the most recent email for a contact but another has a more current job title, the exported record uses the best email and the best job title from whichever source scored highest. Field-level confidence is scored based on recency, source reputation for that field type, and cross-source agreement.

This approach consistently produces better deliverability and data accuracy than single-source or sequential waterfall lookups, where the first available result wins regardless of quality.

Pristine monitors three signal categories natively: job changes (sourced via LinkedIn activity feeds and HR data aggregators), funding events (sourced via Crunchbase and PitchBook API feeds, refreshed daily), and technology stack changes (sourced via BuiltWith and Wappalyzer crawl data, refreshed weekly).

Signals are surfaced inside the prospecting workflow, while the user is building a list, rather than in a separate dashboard. A funding event or job change that surfaces after a prospect list is built goes unactioned by most teams. Pristine shows signals at list-build time so the decision to include or exclude a prospect reflects the most current context available.