The chart is hard to argue with. Impressions flat or climbing, clicks falling off a cliff on the same queries, and an AI Overview now sitting on top of the results where your page used to get the visit. The independent studies put the drop somewhere between half and two thirds of the clicks on queries where an Overview appears, and Overviews now show up on more than half of tracked B2B SaaS keywords. So the traffic is really gone, and the instinct is immediate: figure out how to get it back.
That instinct is the trap. Not because the loss is not real, but because the click was never a single thing, and the panic treats it as one. Before you spend a quarter trying to claw back the number on the chart, it is worth asking which clicks you actually lost, and whether they were the ones that were ever going to matter.
Not all clicks were carrying the same weight
An AI Overview is best at answering the question that has a clean, general answer. “What is a data warehouse.” “How does SSO work.” “What should a QBR cover.” Those are top of funnel, informational, and they were the bulk of the traffic to your explainer content. They were also your lowest-intent visitors: people learning a concept, most of whom were never going to buy from anyone this quarter, let alone from you. The Overview absorbs those most completely, because those are exactly the queries it can fully satisfy in a paragraph.
The queries that feed pipeline behave differently. Branded searches, comparison queries, “best tool for a team like mine,” the bottom-of-funnel questions a buyer asks when they are close to a decision. Overviews are weaker there, because the honest answer is “it depends on you,” and the searcher knows it, so they still click through to judge for themselves. That traffic is more intact than the headline suggests.
So the aggregate “58% CTR drop” is a blended number that hides the part you care about. Most of what you lost was informational volume that looked good in a traffic report and did little for the business. The visits that actually turn into demos were less exposed. If you run a media site that monetizes top-of-funnel volume with ads, this distinction is bad news and you should keep reading elsewhere. If you sell a considered B2B product, you likely lost mostly the clicks that were already a vanity metric, and you are about to reorganize your content strategy around getting them back.
Getting cited is the right move for the wrong reason
The rest of the internet has already arrived at the fix: restructure your content so the Overview cites you instead of burying you, and you win the click back. The advice is not wrong about the tactic. It is wrong about the scoreboard.
When you get cited inside the Overview, the click it sends you is the smallest part of what just happened. The larger part is that your sentence is now the answer the searcher reads, and most of them do not click at all. You are no longer competing for the visit. You are competing to be the source that forms the opinion before any visit. That is a better position than the one you had, and it is worth the work, but if you measure it by the clicks it returns, it will look like a disappointment, because the clicks mostly will not come back. The behavior that changed is not “people click less on your link.” It is “people get their answer without needing your link at all,” and no amount of citation optimization reverses that. It just decides whether the answer they get is yours.
This is the same mistake the whole category makes with AI search generally: pricing the influence by the traffic, when the return arrives off the click. The Overview is just the most visible place it happens. The buyer reads your cited answer, forms a view, and shows up to the sales call already holding it, or does not show up at all because the answer resolved their question. None of that is a line in your analytics. All of it is happening.
The traffic loss is also telling you something useful
There is a second reading of the chart that is more useful than grief. The pages that got hollowed out the hardest were the ones an Overview could fully replace. A definition. A generic how-to. A listicle anyone could assemble from the same ten sources. The Overview did not destroy value on those pages. It revealed how little was there. If a machine can summarize your page into three sentences and cost you nothing a reader misses, the page was commodity, and it was borrowed authority all along.
What the Overview cannot flatten is the content that only makes sense as the whole thing. Your own data that exists nowhere else. A real opinion with a defensible edge. The specific experience of having done the work, with the numbers and the part that went wrong. You cannot summarize a proprietary benchmark into a generic answer, because the value is the specificity. That content survives the Overview, and not by coincidence, it is also the content most likely to get cited, because engines reward the clear, verifiable, specific claim over the vague and general one. The loss is pointing at where your content was thin. The recovery is not winning back the commodity clicks. It is making more of the thing that cannot be commoditized.
What I got wrong
The first time a client’s category filled up with Overviews, I ran it as a traffic-recovery project. We found the pages bleeding clicks, rewrote them answer-first to earn a spot in the Overview, and it half worked: we got cited on several of the queries, and the clicks stayed down. I logged it as a partial failure. The rewrite got us into the answer and the traffic chart did not recover, so by the metric I had chosen, we had not fixed the problem.
Then we looked at the pipeline instead of the chart. The demos were still coming, and the buyers on them were arriving having already read our answer in the Overview, more informed and further along than the cold informational visitors we used to get. The citation had worked exactly as it should. I had been scoring a pipeline win with a traffic ruler and calling it a loss. Now I do not sell Overview work as traffic recovery, because it mostly is not that. I tell clients the click may not come back, and the win is being the source the Overview reads from, measured by whether we are cited and whether the pipeline holds, not by whether the clicks chart returns to where it was.
Why this matters
The traffic report is going to keep looking worse, and the pressure to chase the clicks back will be constant, because the chart is visible and the influence is not. But rebuilding your content strategy to recover informational clicks that were never going to buy is fighting the last war with the last war’s scoreboard. The clicks are not the asset anymore. Being the answer is.
Before you decide the drop is a crisis, segment it by intent and see how much of it you should honestly be sad to lose. Stop optimizing to win back the commodity traffic. Put the effort into the content a summary cannot replace, get it cited so it becomes the answer, and measure whether your buyers are arriving informed by you, not whether a click chart recovered. The number on that chart is the one that is easiest to see and the one that matters least.
If you want to know what your buyers are actually being told in the answers that replaced your clicks, that is the first thing I check in an engagement, and the full method is on the methodology page.