This video shows a quick and easy way to see whether the rollout of Google AI Overviews has affected your traffic. Or read on for the steps in full.
You can’t measure Google AI Overview traffic precisely, but you can see which searches have started sending you fewer visitors since they appeared.
We use Search Console data for this – you need at least 13 months worth of data in your account.
Use Search Console data to diagnose your traffic drop
1. Go to your Search Console account and navigate to the Performance report.
2. Select the Compare option.
3. In the time comparison screen, choose ‘Compare 3 months year over year’.
You’ll see your search queries listed, with those that generate the highest click volumes at the top. Don’t use the overall difference in clicks to diagnose the AIO traffic drop – drops could be due many reasons like 2024 offerings and events which are now closed, and changes in audience demand.
4. Click on the ‘Clicks difference’ column twice – this will show you search queries which have experienced the biggest drop.
5. Search for the queries where clicks have dropped in Google, to see whether an AIO appears in results.
My website hasn’t experienced a significant drop in Google traffic as a result of AIOs, but for this demo let’s pretend that I’d had a significant drop for the query ‘digital performance dashboard’.
An AIO appears for this search. An Ahrefs study showed that the presence of an AI Overview in search results correlated to a 34.5% lower average click-through rate for the top-ranking page compared to keywords without an Overview. Informational content is taking the hardest hit.
What to do next
If you confirm AI Overviews are affecting your traffic:
1. Review your top informational queries:
Is it worth trying to compete with AIOs, or should you make your content richer, more specific, or more visual to add something extra?
Can you create videos for these topics to encourage click-throughs from AIOs, and rank in separate video searches?
Drive traffic from other pages on your site with internal links, instead of relying purely on Google traffic.
2. Prioritise pages where traffic is dropping but conversions or engagement still matter.
3. Consider creating content designed for AI summarisation (structured, factual, authority signals).
4. Test and experiment – there’s no rule book.
Your approach will depend on your goals, your users and the platforms you publish on.
We need a more nuanced approach to our search strategy than just ‘Get more traffic to the website’, which we might have had in the past, and we need to respond to our audience, using data.
What other AI performance data can I get?
- We can’t see accurate data for website mentions or links in AI responses. Paid SEO tools can provide an indication, but the value vs cost equation doesn’t add up for most organisations.
- We can’t see the exact volume of clicks to our website from Google AI overviews, but this filter for GA4 from Dana Di Tomaso gives us a good indication.
- We can see traffic referred to us from other AI engines by creating a custom filter in GA4 and updating it when new engines appear.