At some point this month somebody is going to ask you why traffic is down, and the true answer sounds exactly like an excuse.
Rankings held. Nothing was penalised. The content is the same content that worked last year. Google simply started answering the question on its own page, and the click you used to get went to nobody.
That explanation is correct and it will not survive a review meeting, so here are the numbers to put behind it, and then the argument I actually want to make, which is that you should stop reporting sessions as a headline number at all.
The numbers, so it stops sounding like an excuse
SparkToro’s analysis puts zero-click Google searches at about 68% in early 2026, up from roughly 60% two years earlier. Fewer than one search in three now sends a click anywhere.
Ahrefs measured the presence of an AI Overview against a comparable query without one and found the top ranking page taking around 58% lower click-through when the overview was there. When they first ran that analysis the gap was 34.5%, so it widened considerably in a year. Pew’s work put the relative decline nearer 47% and Seer Interactive around 61%, which tells you the exact figure depends on method, and that every method agrees on the direction.
Queries that trigger an overview show a zero-click rate around 83%, against roughly 60% for queries that do not.
One honest caveat, and it is SparkToro’s own. Zero-click searching was rising before AI Overviews existed. Featured snippets and knowledge panels started this a decade ago. The overviews did not create the trend, they accelerated it, and pretending 2026 is a sudden ambush makes you sound like somebody who was not paying attention in 2023.
The part I would argue about
Here is the claim. A number that falls when you are doing badly and also falls when you are doing well is not a metric. It is a thermometer in a room where the temperature no longer relates to anything you control.
Sessions is now that number.
Your traffic can drop 40% because your content got worse. It can also drop 40% because your content became the answer that Google reads out, which is the single strongest signal available that you wrote something authoritative. Both produce the same line on the same chart, and no amount of staring at that chart separates them.
So reporting sessions as the headline is worse than useless in 2026. It is actively misleading, and it drives teams towards the exact wrong response, which is publishing more of the informational content that just stopped earning clicks.
If your monthly report leads with traffic, you have a reporting problem before you have a search problem.
What to report instead
Four things. None of them require new software.
Impressions against clicks, tracked as a ratio. Search Console gives you both. Impressions holding or rising while clicks fall is the signature of being read rather than visited, and it means you are still winning the query. Impressions falling is the one that should worry you, because that is genuine lost visibility rather than a redirected click.
Most people look at clicks alone and conclude the sky is falling. The two lines side by side tell you which of the two situations you are in, and it takes ten minutes to set up.
Branded search volume. This is the closest thing to a replacement for the click you lost. Somebody reads an AI answer that cites you, does not click, and then two days later searches your company name. That sequence is now a normal path to a customer, and it shows up as branded search growth with no attributable source.
Track it monthly. If branded search is rising while total traffic falls, your content is working and the measurement is the thing that broke.
Conversion rate of what still arrives. The traffic that survives is more commercial than what you lost, because informational queries were hit hardest and transactional and local queries barely moved. So your remaining traffic should convert better. If sessions fell 40% and enquiries fell 5%, that is a good quarter dressed up as a bad one.
Enquiries. The number that was always the point, and the only one on this list your business would notice if it went to zero.
The trap, and it is a very Indian one
The instinct when informational traffic falls is to publish more informational content, faster, because that is the machinery you already built.
A lot of Indian SMEs and agencies built an entire SEO practice on “what is”, “how to” and “top 10” blog posts, because they were cheap to produce at scale and they used to work. That inventory is the exact inventory the overviews absorbed. Producing more of it in 2026 is paying to fill a bucket with a hole in the bottom.
I split this argument out properly in whether SEO is still worth it for a small business, and the short version is that the informational half took the damage and the commercial half is close to untouched.
For local businesses here the picture is better than the headlines suggest. Somebody searching for a clinic in Indiranagar or a chartered accountant in Andheri is still clicking, still calling, still opening Maps. That traffic did not go anywhere, which is why the local search work has quietly become the highest return SEO most Indian service businesses can do.
The uncomfortable version
Some of the traffic you lost was never worth anything.
If a page brought two thousand visits a month for years and produced no enquiries, its disappearance costs you exactly nothing, and the panic about it is really grief for a number that felt like progress.
That is why this landed so unevenly. Teams whose content was doing commercial work barely noticed. Teams whose content was doing volume work lost the thing that had been standing in for results, and are now discovering that it was standing in for results.
Neither group changed what they were doing. One of them just found out.
What I would change on Monday
Move sessions out of the first slide. Not out of the report, out of the first slide.
Put impressions, branded search volume and enquiries above it, in that order, with a single line explaining why. Then go back through your last two years of content and mark each page as informational or commercial. If the informational pile is most of it, that is the finding, and it decides your next two quarters more than any tactic will.
My notes on AI search go into what actually earns a citation, which is the other half of this.
If you’d like to talk through what your own numbers are really saying, drop me a line on email, WhatsApp or LinkedIn and we can have a quick chat. I’m contracted full time so this isn’t a pitch.