A Practical Guide to Multi-Source Attribution in the Age of AI SEO
A Practical Guide to Multi-Source Attribution in the Age of AI SEO
With AI overviews and zero-click searches reshaping how people discover your business, a single analytics dashboard is no longer enough to understand what's actually driving growth. We've put together a practical framework for combining referral traffic, conversion data, and on-site analytics so you can make confident decisions, not educated guesses.
Why Attribution Is Harder Than It Used to Be
A few years ago, the journey was relatively straightforward. Someone searched, clicked, landed on your site, and converted. Your analytics told you the story. Today, that story has more gaps in it than ever before.
AI overviews now answer questions directly in search results. Users get the information they need without visiting your site at all. Zero-click searches are no longer an edge case, they're a significant chunk of how people interact with your brand online. As Cyrus Shepard and others in the SEO community have pointed out, AI SEO attribution is still a major challenge, precisely because the top of the funnel has become far less visible. You can be doing everything right and still see flat organic traffic numbers, because the people you're reaching never needed to click.
That shift demands a different way of thinking about what counts as a signal.
Stop Obsessing Over Individual Sessions You Can't Pin Down
One of the most common frustrations we hear from clients is the volume of traffic showing up as 'direct' or '(none)' in their analytics. It feels like missing information. But we've found that obsessing over individual sessions you can't attribute is largely a waste of energy.
Our approach is to look at patterns across a rolling 90-day window instead. When you zoom out, trends become visible that single sessions never reveal. A steady rise in branded searches alongside a spike in direct traffic often points to a podcast appearance, a PR mention, or a piece of content that travelled further than your UTM tags did. The pattern tells you something. The individual session usually doesn't.
That said, missing data is information too. Before you write off a column full of blanks, ask why those values are blank. A sudden jump in unattributed direct traffic might mean a dark social share is doing the rounds, or that a campaign went out without proper tagging. Dropping those rows without investigating them first means losing a signal that could actually matter.
UTM Discipline Is Non-Negotiable
The biggest crack in the attribution picture is almost always UTM discipline, or the lack of it. We've made consistent tagging non-negotiable across every campaign and partner link we run, because without it you're essentially building a jigsaw with half the pieces missing.
This sounds obvious. It's also the thing most agencies let slip. A single untagged newsletter link, a partner posting your URL without parameters, a paid social ad where someone copy-pasted the landing page URL rather than building it properly, each one quietly corrupts your data. Over time, those gaps compound.
Our fix is simple: a shared UTM naming convention, documented and enforced, that every team member and external partner uses. Source, medium, campaign, consistent every time. It takes an hour to set up properly and saves countless hours of guesswork later.
When the Top of the Funnel Goes Dark, Follow the Bottom
If AI overviews mean fewer people are clicking through to your site, top-of-funnel metrics like impressions and organic sessions become harder to trust as indicators of whether your marketing is working. So we've shifted weight towards conversion data from the bottom of the funnel.
When top-of-funnel signals get noisier and harder to attribute, the clearest proof that something is working is whether it's actually producing clients. That means tracking form submissions, phone calls, proposal requests, and actual revenue with the same rigour you'd apply to traffic. It means making sure your CRM and your analytics platform are talking to each other. And it means being willing to treat a steady stream of good-quality leads as a positive signal, even if you can't perfectly trace every one back to its origin.
This is not an excuse to stop caring about attribution. It's a recalibration. You still want to understand where your best clients are coming from. You just need to be honest about which signals are reliable and which ones have become genuinely noisy.
Bringing It Together: A Practical Framework
Use a rolling 90-day view. Patterns across time are more useful than daily or weekly snapshots when attribution is fragmented.
Investigate blank values before discarding them. Missing data often points to a gap in your tagging or a channel you haven't accounted for.
Enforce UTM consistency across every channel and partner. Document it, share it, and check it regularly.
Weight your KPIs towards bottom-of-funnel conversion data. Leads, sales, and client value are harder to fake than traffic numbers.
Cross-reference your tools. GA4, Search Console, your CRM, and any call tracking software should all be in the same conversation, not siloed into separate reports.
You Don't Need Perfect Data. You Need a Consistent Method.
No attribution model is perfect. Some sources will always be murky, and that's fine. The goal is not a flawless picture, it's a consistent method that you apply over time, so that the gaps stay the same size and the trends you do see are meaningful.
If you're finding that your current reporting raises more questions than it answers, we can help. At Air 66, we work with businesses across the UK to build attribution frameworks that actually reflect how customers find and choose them, not just how analytics platforms prefer to count things. Get in touch and let's talk through what's working and what isn't.