Sixty-three percent of what drives a brand's visibility in AI answers comes from long-term brand equity built up over years, not from anything done on or for AI platforms this quarter.
Dan Wilson at Charlie Oscar got that number from a regression analysis across 30 brands and 6 categories, built specifically with lagged variables to separate what's causing visibility from what merely correlates with it. Citation volume, the thing most AEO advice is about, accounts for 11%.
Brand equity explains most of it, not citations
Wilson's breakdown splits cleanly: 63% long-term brand equity, 22% total marketing spend, 11% citation volume, 4% influencer and PR reach. The 63% is described as "a strong baseline visibility driven by long-term brand equity" - the accumulated effect of a brand having been talked about, reviewed, and searched for over years, which is exactly the type of signal a language model trained on the internet would pick up on. The remaining 37% is what a marketing team can move in the short term, and citation-building work, the actual substance of most AEO advice, is the smallest piece of that.
The correlation data doesn't say what it's often made to say
Ahrefs' own 75,000-brand study of AI Overview visibility is frequently cited as proof that branded search volume drives AI visibility. Check the actual numbers and it's the third-strongest factor, not the first: branded search volume correlates at 0.392, but branded web mentions come in at 0.664 and brand-name anchor text at 0.527. Backlinks, at 0.218, barely register. Paid advertising moves it even less - branded ad traffic and ad cost both sit at roughly 0.216. Ahrefs' own conclusion from that data: "Ads won't save you if you want AI Overview visibility." Their closing line goes further, and it's the same conclusion Wilson's regression reaches independently: the struggle for AI visibility "still looks a lot like brand building - because that's exactly what it is."
Even where it converts, the business impact is still small
Wilson's analysis didn't stop at what drives visibility. It also modelled what that visibility is actually worth. LLM-tracked sessions made up under 1% of total tracked sessions across the brands studied. At a category peak, LLM visibility explained up to 5% of incremental demand - smaller than the short-term impact of paid marketing (29%) or a rise in branded search (20%) in the same model. Of that 5%, only the citation-driven share (11%) is attributable to AEO-style tactics specifically. Run the arithmetic and it comes to a net 0.55% of sales volume from getting more citations. Wilson's own summary: "Do everything right, and it will barely make a difference to your next two quarters of results."
The traffic-quality numbers move faster than the story around them
Adobe's AI-referral conversion data is real, and it has moved a long way. In July 2024, AI-referred traffic was 43% less likely to convert than other channels. By February 2025, that gap had closed to just 9%. A year later, it had flipped outright: Adobe's Q1 2026 data shows AI-referred traffic converting 42% better than non-AI traffic, a full reversal from converting 38% worse twelve months earlier. A metric that swings that far in a year is a live, correlational signal, not settled proof that any particular AEO tactic caused the swing.
Marketing has been here before with paid brand search. Haus's large-scale analysis of real branded search experiments found just under a 10% average lift for tests that measured a post-treatment window, and even that varied by competitive pressure - from 9% in low-competition categories up to 37% in high-competition ones. P&G ran an experiment cutting brand search spend and saw no change in business outcomes. eBay paused brand search ads for 30% of its users and found no measurable change in sales. Platforms have a financial incentive to let a channel look more valuable than it is. Track whether your own visibility is actually moving. Don't assume a specific tactic caused whatever number you're looking at this month.
A few common follow-up questions
Does this mean AEO work is a waste of time?
No. Wilson's own view, not this piece's: the current business impact is small, but the trend line is rising, and it may be far more consequential by 2027. It's the smaller lever, not a useless one.
If citations only explain 11% of visibility, why does everyone talk about them?
Citations are the visible, trackable part - the specific thing most AI visibility tools are built to measure. Brand equity built over years shows up in the data as a stable baseline, not an event, so it's easy to overlook in favour of whatever moved this week.
Where should the marketing budget actually go?
Mostly where it already goes: broad brand-building and existing marketing spend, which this research puts at 85% of what drives AI visibility between them. Citation work is worth doing. It's not a substitute for that.