Skip to content
How to measure SEO performance and turn search data into commercial decisions
Plenty of board meetings now have a version of the same slide. Organic traffic is down, AI Overviews answer questions before anyone reaches the website, and someone asks if search is still worth the money. Some websites are losing clicks, mostly on queries Google now answers itself, so it’s easy to think search data no longer helps with commercial decisions.
However, people still search for the problem, the product and the competitor’s name. That record of demand is still useful, even though it misses what happens inside AI assistants. Traffic alone has become a less reliable sign of demand, so judging search by visits can lead a business to cut one of its best demand signals. To me, that’s the wrong thing to cut.
That’s the gap our SEO work is built around. We treat search as a sign of demand first and a source of traffic second. Then we connect organic and AI-referred visits to sales and other business results.
Where the "traffic equals value" idea comes from
Search reports grew up around two numbers, where a page ranked and how many visits it got. For years the two moved together. A higher position brought more clicks, so visits became a shortcut for business value.
That link is weaker now, as an Ahrefs study of 300,000 keywords shows. It looked at desktop click-through rate, which is the share of searchers who click a result. For queries with an AI Overview, the top result’s rate fell from 7.3% in December 2023 to 1.6% in December 2025. That’s 58% below the rate Ahrefs estimated if AI Overviews had not appeared.
The study is observational and Ahrefs sells SEO software. Even so, it points to a large change in how people click within the sample. My view is that a ranking report on its own now tells a board very little.
How AI Overviews and answer engines changed search
In May 2025, a year after AI Overviews launched and changed who gives the answer, the SEO platform BrightEdge reported that Google search impressions were up 49%. Over the same period, click-through rate fell 30%.
Impressions count how often a page appears, which is different from how many people searched. A lower click-through rate also doesn’t always mean fewer clicks in total. It does mean each appearance turns into a visit less often.
How often AI Overviews appear varies. Ofcom’s Online Nation report, published in December 2025, put them on about 30% of UK searches. Semrush found them on 24.61% of its tracked keywords at their peak in July 2025 and on 15.69% by November. Those figures are shares of a sample of more than 10 million keywords, so they don’t describe every search.
Semrush also tracked keywords before and after an AI Overview appeared. The share of searches that ended with no click at all fell slightly, from 33.75% to 31.53%. That measures whether a search led to any click, so it doesn’t clash with the Ahrefs figure for the top result.
This is the space answer engine optimisation (AEO) covers. AEO means earning a place inside the answer, across AI Overviews, Google’s AI Mode and assistants such as ChatGPT. Organic ranking still matters, but a second contest now runs alongside it.
How to measure SEO performance in AI search
The basics still apply, so indexing, rankings for key queries and sales from organic search remain the foundation. On top of that, we add three layers, because each one answers a different business question.
Demand – whether interest in the category and the brand is growing. Search shows this whether or not anyone clicks, though it misses demand inside AI assistants. Changes in search habits, data coverage and platform use can also shift what you see from one month to the next.
Presence – whether the brand appears on the page and in the answer. An AI Overview citation and a brand mention without a link are different results, so track them separately.
In an Ahrefs study of 75,000 brands, mentions of a brand across the web had the strongest link with AI Overview visibility. Branded search volume mattered too. That doesn’t prove more branded searches cause more AI mentions, though.
Value – what organic and AI-referred visits are worth. Referrals from AI assistants belong here, along with sales and revenue. In a July 2025 study, Semrush reported a 4.4 times higher conversion rate for tracked AI-referred visitors than for traditional organic visitors.
Conversion rate alone doesn’t show revenue, profit or customer lifetime value, and Semrush studied its own sector, so treat the figure as a rough guide.
Which commercial decisions search data should drive
Demand planning and brand tracking can use signals that don’t depend on clicks. Investment decisions also need sales and profit data, as these examples show.
Demand planning – Google Trends shows interest as a score from 0 to 100, based on a sample of searches, as Google’s own guide explains. That makes it useful for spotting direction and seasonal patterns, but not for counting searches.
Brand health – share of search divides a brand’s search volume by the total for a set group of brands. The IPA calls share of search a strong proxy for the demand side of market share. It also notes that the link between the two varies by category and country.
Our own SEO work with Nature’s Best shows why investment decisions need sales data as well as traffic. Their search traffic was falling after they removed a section of the site that brought in low-quality visits and rarely converted. We fixed the technical issues, cut or combined pages that weren’t earning visibility, and trained the in-house team to brief and write content for search. Over 10 months, organic revenue rose 65.5% year on year and organic conversion rose 67.1%.
Now that AI Overviews are part of the picture, judging search by revenue rather than visits matters even more. In this illustrative scenario, with hypothetical figures throughout, a retailer’s organic visits fall 15% year on year and the board wants to cut the SEO budget.
Before deciding, the team checks the business results and finds organic revenue is down only 2%, while gross profit from organic orders is flat. Conversion rate has also improved. The orders that remain carry a slightly higher gross margin.
Breaking down the lost visits shows where the drop is concentrated. Almost all came from broad, informational queries that now show AI Overviews and rarely led to orders. Transactional queries kept their rankings and revenue. AI Overviews help explain the pattern, though they may not be the only cause.
Category demand is flat, while the brand’s share of search across itself and four tracked competitors has risen from 18% to 21%. So the brand is losing low-value clicks while maintaining its commercial performance.
The team keeps the overall budget and tests a new split, with product, comparison and delivery pages first and a smaller share for improving and measuring AI visibility. Informational content keeps a role, since it can still help people discover the brand and buy later. That’s the kind of check I’d want every board to make before cutting SEO spend.
How to turn search data into decisions
Each step covers a blind spot in the one before, so the order matters.
- Check demand signals – treat search interest and share of search as signs, and check them against sales, enquiries, customer records and customer research.
- Fix the competitor set – report share within the tracked brand set, using the same brands each month.
- Compare like with like – compare click-through rate for tracked queries with and without AI Overviews, allowing for changes in rankings, device, market and search intent.
- Track AI results separately – log AI Overview citations and brand mentions with a tracker such as Semrush or Ahrefs Brand Radar, and report assistant referrals with your other visits.
- Value each visit – measure revenue per organic visit, and focus full service CRO work on pages where visitors need to explore, evaluate or take action.
- Invest in pages with a clear reason to visit – AI can sum up comparisons, prices and local details too, so fund what a summary can’t give, such as tailored calculations, live stock, demos or a way to buy.
What should the board see from search?
A board deciding where to spend needs to know if category demand is rising. It also needs to know if the brand has gained share over the last 12 months and what customers from search are worth. Traffic sits underneath as a supporting measure.
If your search report stopped counting visits tomorrow, what would it tell the board about next year’s demand? If you’d like a second view on that, get in touch with our team.
Frequently asked questions
How do you measure SEO performance?
Start with the basics, which are indexing, rankings for key queries and sales from organic search. Then add three layers. Demand covers search interest and share of search, presence covers AI Overview citations and brand mentions, and value covers organic and assistant referrals, conversions and revenue.
What is AEO?
AEO, or answer engine optimisation, means earning a place inside AI-generated answers, such as Google’s AI Overviews or ChatGPT. It builds on SEO, since Google says a page must be indexed and able to show with a snippet to appear as a supporting link in AI Overviews or AI Mode.
What is the difference between GEO and SEO?
SEO (search engine optimisation) aims to rank pages in search results. GEO (generative engine optimisation) aims to get content cited or summed up by AI tools. SEO is measured through visibility, traffic and business results, and GEO adds AI mentions and citations.
How do you measure SEO ROI?
SEO ROI, or return on investment, is the return from organic search minus the programme cost, divided by the programme cost. Use incremental profit, the extra profit SEO brings in, where you can measure it, because attributed revenue alone is not ROI.
Sources
Ofcom: From apps to AI search, how the UK goes online in 2025 (Online Nation 2025, published December 2025)
Ahrefs: AI Overviews reduce clicks by 58% (February 2026 update, 300,000 keywords, desktop CTR)
BrightEdge: AI Overviews one year review (May 2025)
Search Engine Land: New Google AI Overviews data, search clicks fell 30% in last year (BrightEdge, May 2025)
Semrush: AI Overviews study (10M+ keywords, January to November 2025)
Ahrefs: An analysis of AI Overview brand visibility factors (75,000 brands, May 2025)
Semrush: AI search and SEO traffic study (July 2025)
IPA: Share of search (October 2021)
