Read visitor journeys and attribution
Investigate how people arrive, which pages they visit, and what attribution evidence actually supports.
Journeys & Pulse connects site behavior with known people when identity is available. It helps Northgate ESD Packaging understand whether visitors reach useful content and quote forms, while keeping observed traffic separate from inferred attribution.
Start with a consistent time window
- Select the Northgate demo account.
- Open Journeys & Pulse > Overview.
- Check the displayed date range.
- Read Sessions Over Time and Traffic Sources together.
- Inspect Conversion Paths for a relevant entry page.
A conversion in these paths is a session containing a form submission. It is not a count of sales. Compare the session count and conversion rate together: a page with one conversion from two sessions has a high rate but little evidence of repeatable performance.
For Northgate, compare visitors entering through packaging guidance with visitors entering directly on the quote page. Use the comparison to ask a useful question, such as whether guidance helps buyers reach a form. Do not assume the entry page caused the conversion merely because the two appear in the same path.
Investigate visitors and known people
- Open Visitors.
- Filter by a relevant page path when investigating a specific page.
- Inspect recent activity and whether a visitor is known.
- Open People when you want to review an identified person's journey.
- Read the person's event history alongside any available Pulse information.
Visitors represents browser activity; People represents identified records. Several events do not necessarily represent several people. An anonymous visitor is not automatically a contact, and a person can have more than one associated visit.
When an identity becomes known, associated earlier activity can appear in the person's journey. That history adds context for a conversation, but the CRM remains the contact system of record. Use it for the authoritative contact details and normal follow-up process.
Interpret AI attribution carefully
AI-Referred Traffic separates three kinds of evidence. Confirmed traffic has observed source evidence. Self-reported attribution comes from a person's answer about how they heard of the business. Suspected traffic is an estimate based on behavior consistent with a referral that lost its referrer.
Do not add suspected traffic to confirmed traffic and label the result confirmed. A direct landing deep in a site may be consistent with a copied AI recommendation, but it can have other explanations. Self-reported answers add another perspective rather than repairing every unknown visit.
The Citation Observatory answers a different question: whether a model response mentioned the brand or returned a source URL in a recorded run. A citation observation does not prove that someone visited the site. Journeys is where you inspect the visits that were actually tracked.
Turn a finding into a next step
Suppose Northgate's guide receives visits but few form submissions. Inspect Pages for engagement context, then inspect the relevant form in Forms or Signals. The next action might be reviewing the quote call to action or checking form delivery, depending on what the evidence shows.
Write down the period, page, counts, and uncertainty. “The guide had visits but no tracked submissions in this period” is useful. “Nobody wanted this product” goes beyond the evidence.
If data is missing
Check the account and time range first. No tracked activity may reflect site setup, tracking availability, or a genuinely quiet period. An unnamed visitor may simply not have identified themselves. Missing Pulse data does not mean a score of zero.
Ask the responsible operator to check the tracked site configuration in Settings if expected visits never appear. Keep one known test path and time available so they can investigate a concrete observation.