Skip to main content
Once a session is captured, Swan works through a pipeline to figure out who was behind it and whether it’s worth surfacing to your trigger.

​
How identification works

Every Swan tracking script automatically includes two things, with no separate setup required: the RB2B pixel, and a pixel that feeds Swan’s own matching layer. Swan tries to resolve each session through a few paths, roughly in order of speed:
  • RB2B — identifies visitors browser-side and reports matches back to Swan through a webhook as they happen, so RB2B-sourced person-level matches are the closest thing to real time. The trigger itself fires once that visitor’s session goes quiet, which is usually within about half an hour.
  • Swan’s own matching layer — an internal data process that resolves visitor IP addresses to companies and, where enough signal is available (for example a hashed work email captured during the visit), to a specific person’s LinkedIn profile. This runs on a fast cycle, roughly every 25 minutes, independent of RB2B.
  • A general company-level fallback — for sessions neither of the above resolved, Swan waits for the session to go quiet (about 2 hours of inactivity) and makes one more attempt to match the visitor’s IP address to a company using Swan’s own IP-to-company data. This is the slowest path and accounts for most of the “up to a couple of hours” delay.
Once a visitor resolves to a LinkedIn profile — whether through RB2B or Swan’s own matching layer — Swan runs a short waterfall across its own data providers to fill in the rest of the contact: name, title, and company details. If your identification mode — set under “Who should we identify?” on the tracking script’s Configuration → Settings tab — is set to only match known contacts, Swan skips the slow company-level IP fallback and won’t invent a company-only match from that path. Note, however, that company-level identifications arriving in real time from RB2B (and similar provider pixels) aren’t currently filtered by this setting and can still fire.

​
Qualification pipeline

Not every identified visit reaches your trigger. Before a trigger fires, Swan filters out:
  • Visits with no resolvable company domain or person data — if Swan can’t tie the visit to a usable company domain or a specific known person, the visit is dropped rather than fired with incomplete data.
  • Companies on Swan’s internal exclusion list — a small built-in blocklist applied across all customers, separate from anything you configure.
  • Excluded companies from your CRM — companies belonging to a CRM list you’ve marked as excluded (for example, existing customers).
  • Near-duplicate matches — if the same company (by domain or LinkedIn URL) was already matched within the last 20 minutes, the newer match is treated as a duplicate and doesn’t create a second trigger event.
  • The daily identification limit — once your org hits its daily cap (from the tracking script settings), Swan stops serving the tracking script for the rest of the day, so no further visits — person or company — get identified until the cap resets.

​
Cost

  • Person-level identification: 2 credits.
  • Company-level identification: 1 credit.
  • RB2B-sourced identifications: 0 credits.
These are charged when a visit clears the qualification pipeline and fires your trigger — see credits.

​
Privacy

Identifying a visitor involves processing their IP address and, where a match is found, the resulting person or company identity. Make sure your site’s privacy policy accounts for this kind of visitor identification, and that you’re comfortable with your legal position in any jurisdiction your traffic comes from before enabling it.

​
Common questions

Why does identification take up to two hours instead of being instant? Most visits resolve faster than that — RB2B matches arrive close to real time, and Swan’s own matching layer runs on a roughly 25-minute cycle. The two-hour case is a fallback: Swan waits for a session to go quiet before making one more company-level match attempt on anything that wasn’t already identified. What percentage of visitors can Swan actually identify? It varies by traffic source, industry, and how much of your traffic is from company networks versus residential ISPs — there’s no fixed identification rate. If the same company visits again next week, do I get notified again? Yes — Swan currently only treats matches within about 20 minutes of each other as duplicates, so a visit from the same company a day or a week later creates a new trigger event. Can I see visits that didn’t qualify, not just the ones that fired my trigger? Ask Swan directly, or check your trigger’s execution history at its Logs tab (https://agent.getswan.com/triggers/{triggerId}/logs) — unqualified visits aren’t shown as a standalone visitor list, but the underlying session data isn’t discarded either. Does Swan use cookies to identify visitors? Identification relies primarily on the visitor’s IP address and pixel-based signals (including RB2B’s), not on third-party ad cookies.