Curious what others are seeing. On a few small local business sites I look after, impressions in Search Console went up this year but clicks stayed flat or dropped a bit, mostly on "how to" and "cost of" type queries. Branded and "near me" searches still convert fine. Are you changing anything because of this, like moving effort from blog posts to service pages, adding more FAQ style answers, or just focusing on the Google Business Profile? Would like to hear what is working for you.
Seeing the same pattern on informational pages: impressions up, clicks flat, mostly on queries where the answer fits in a paragraph. "Cost of" is the one I'd watch most, because a generic price range is exactly what the overview gives, and the person who still clicks usually wants a real local number. What has helped on small sites: - Service pages that answer price and process questions directly, with a range for the area and what pushes it up or down. That is something an overview can't build from generic sources. - Keeping the how-to posts that bring calls or links, and folding the thin ones into the service page they support instead of writing more of them. - Tracking calls and form fills by landing page, not just clicks. A drop on "how to" queries often barely touches leads, since most of those visitors were never going to hire anyone. Agree on the GBP point. For local, that is where a lot of the actual decisions get made now.
I’d look at Search Console and leads together. If a page is getting impressions but fewer clicks, check whether calls and forms from that page changed too. If service pages still bring leads, I’d keep the useful how to posts and make those pages answer practical questions about local prices and what happens next. Comparing with the same time last year can help rule out seasonality.
I'd stop judging those pages by clicks and look at what the visitors who still arrive do. Set up Search Console filters for your informational queries and compare them against branded and "near me" queries, then check if calls or form fills tied to those pages moved at all. If the leads are flat, the lost clicks were mostly people who only wanted a quick answer, and I'd put the effort into service pages with local examples, photos of finished jobs and a clear next step. Where a how-to post does earn links or calls, update it with a local detail and an obvious contact option instead of writing new ones.
Yes, that split is exactly what I would expect, and it fits how Google describes AI Overviews: they are meant to give people the gist of a complicated question quickly, with links as a jumping-off point to learn more. So a short how-to or a generic cost answer can be satisfied on the results page, while a branded or near me search still needs a real business behind it. Before moving effort around, I would split the Search Console data by query intent instead of looking at the site total. If the drop is mainly informational, I would not judge those pages only on clicks. I would check whether the people who still click go on to call or fill in a form, and whether the page is actually being shown as a supporting link in the overview. For the pages losing clicks, I would add the local detail an overview cannot give generically: the price range for that area, what affects it, how long the job usually takes, and what happens next if someone books. That gives the searcher a reason to click through for the specifics. I would keep the how-to posts that still bring enquiries and fold the thin ones into the relevant service page, rather than deleting useful content outright.
I'm seeing the same thing on a couple of local sites I look after, impressions climbing while clicks on the how-to pages just sit there. What helped me stop worrying about it was splitting the queries in Search Console instead of looking at the site total, because the drop is almost all on the informational ones and those visitors were mostly after a quick answer anyway. The people who still click through on a cost query usually want a real number for their area, so I put the local price range and what changes it right on the service page now instead of burying it in a blog post. One thing I would add is to watch the GBP too, a lot of people ring straight from there these days without ever touching the site, so the clicks on their own make it look worse than it is.
One annoying thing with that, Search Console doesn't split the AI Overview stuff out at all, Google just counts it inside the normal Web numbers. So you can't really prove it's the overviews, you're kind of guessing from which queries dropped.
True, there's no AI Overview filter in Search Console, so it's always an inference. You can make it a fairly solid one, though. Take the queries that lost clicks and search them yourself, logged out and set to the right location, and note which ones show an overview. Then compare CTR before and after for queries sitting at a similar average position. If the overview queries dropped and the ones without an overview at the same position held steady, that's about as close to proof as that data will get you. Some rank trackers also record whether an overview showed for each tracked keyword, which saves checking by hand. Worth repeating the check every month or so, because the overviews come and go on the same query.
Same pattern on the small sites I look after — the click loss is almost entirely on top-of-funnel 'how to' queries. The money queries (service + location, price, 'near me') have held up much better, so the split by intent that others mentioned above is the right way to analyse it. Two things that have actually helped: 1. Treat AI Overview presence as a content filter. Before writing a new informational post, I search the target query myself and check whether an overview already answers it well. If it does, that page is a bad traffic bet — I either skip it or fold the topic into a service page instead. The queries without strong overviews (comparisons, prices, local specifics) are where the clicks still are. 2. Format for citations rather than clicks. Short definitions, price ranges, comparison tables and clear step lists get picked up in overviews more often. You don't win the click, but you do get seen as the source, and I've seen a small lift in branded searches after that. And yes, calls and form fills per landing page is the metric that matters now — clicks alone will mislead you on these sites.
Yes, informational queries tend to lose clicks when an AI Overview appears, while local and transactional searches are usually much less affected, which fits your "how to" losses and branded/"near me" searches still converting. For those small business sites, I would split Search Console queries by intent and compare the same period year over year rather than move all the blog effort to service pages. As AmberEcho730 and aiangels.io noted, there is no dedicated Overview filter in Search Console, so use manual SERP checks or a tracker to group queries that trigger Overviews versus ones that do not, allowing for location and ranking changes. Look at impressions and CTR together, not just clicks, and keep calls and form fills alongside that comparison. Being cited inside the Overview matters too, though a citation does not guarantee a visit. Keep Googlebot unblocked and the key answers in plain HTML rather than only JS, and check Google-Extended separately for your intended AI permissions, since it is not the control for appearing in Google Search Overviews. I would add short FAQ-style answers about local cost ranges and what changes the price to the relevant service pages, with appropriate schema matching the visible content, without expecting markup alone to earn a citation. There is also a free no-signup AI visibility audit that checks whether ChatGPT, Perplexity and AI Overviews can actually read and cite a page: https://quanticdata.io/tools/ai-visibility-audit/ — disclosure: I'm on the team that built it.