AI referrals broke your attribution quietly: AI assistants send buyers with no UTM and often no referrer at all. Those orders land in direct, which is already the bucket that absorbs everything nobody can explain.
Read the full article below for detailed insights and actionable strategies.
The attribution problem
One sale. Four channels. 400% credit claimed.
Reported revenue: €400 · Actual revenue: €100 · Gap: €300
A new channel arrived without a column in your report. Assistant referrals mostly land in direct, and direct is already where every unexplained order goes.
Why the channel has no column
A visitor who asks an assistant a question and follows a link to your store often arrives with no UTM, no click id, and sometimes no referrer header at all. Some assistants strip the referrer. Some answers are read and the domain typed by hand. Some links are copied into another app first.
None of that is malicious and none of it is new in kind. It is the same mechanism that has always made direct the largest and least useful row in the report.
What that does to the numbers
| What you see | What is actually happening |
|---|---|
| Direct grows, no campaign explains it | Assistant referrals arriving without identifiers |
| Branded search rises with no spend change | Demand created upstream, harvested by brand terms |
| A referral row for one assistant looks small | It counts only the sessions that kept a referrer |
The third line is the trap. A visible assistant referral row invites you to conclude the channel is small, when what you are looking at is the fraction that happened to be tagged.
The question worth asking
Not "how much traffic do assistants send." That number is unknowable from your analytics alone, and chasing it produces a confident figure built on the sessions that were easiest to count.
The question is whether those visits are incremental: would the order have happened anyway through branded search, or did the assistant create demand that did not exist. Those two cases look identical in a session report and imply opposite budget decisions.
What you can do this week
Segment direct by landing page. Assistant referrals tend to land deep, on a specific answer or product page, not on the homepage. A step change in deep direct landings is a real signal even when the referrer is gone.
Then stop treating the result as a measurement and start treating it as a hypothesis. The difference between a correlation and a cause is the whole subject here.
Where a causal read fits
A read on a Google Analytics export returns, per channel, an estimate with a confidence interval, a coverage share, and an honest label for the channels too small to resolve. It is 99 euro once, refunded if it does not move a budget decision. The channels arriving without identifiers are exactly the ones that get labelled rather than scored, which is the correct answer rather than a flattering one.
The interactive demo shows the fields that read depends on, with no signup.
Related answers
Key Terms in This Article
Attribution
Attribution identifies user actions that contribute to a desired outcome and assigns value to each. It reveals which marketing touchpoints drive conversions.
Attribution Report
Attribution Report shows which touchpoints or channels receive credit for a conversion. It identifies which campaigns drive desired actions.
Confidence Interval
Confidence Interval is a statistical range of values that likely contains the true value of a metric. In marketing analytics, it quantifies uncertainty around estimates, indicating the precision of an outcome or causal effect.
Correlation
Correlation is a statistical measure showing a relationship between variables; it does not imply causation.
Google Analytics
Google Analytics is a web analytics service that tracks and reports website traffic.
Incrementality
Incrementality measures the true causal impact of a marketing campaign. It quantifies the additional conversions or revenue directly from that activity.
Landing Page
Landing Page: A single web page that appears after clicking a search result, marketing promotion, email, or online advertisement.
Product Page
Product Page is a webpage dedicated to a single product. It includes images, descriptions, pricing, and purchase options.
Related Articles
Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.
Ready to see your real numbers?
Own the budget? Upload your GA4 export and see which channels drive incremental sales, with confidence intervals, in minutes. Have to defend it? Start with the live demo and take the read to your CFO.
Full refund if you don't see value.
Stay ahead of the attribution curve
Weekly insights on marketing attribution, incrementality testing, and data-driven growth. Written for the person who owns the budget and the person who has to defend it.
No spam. Unsubscribe anytime. We respect your data.
Frequently Asked Questions
Why does AI assistant traffic show up as direct?
Many assistant referrals arrive with no UTM parameters, no click id, and sometimes no referrer header, because the assistant strips it or the visitor copies the URL. With no identifier, analytics has nothing to classify the session as, so it falls into direct.
Can I measure how much traffic AI assistants send me?
Not precisely, from analytics alone. A referral row for an assistant counts only the sessions that kept a referrer, so it understates by an unknown amount. Treat it as a floor, not a measurement.
Is AI referred traffic incremental?
That is the question worth answering and it cannot be read off a session report. The order may have happened anyway through branded search. Separating the two needs a causal design, not a bigger referral table.