A tutorial with 200 views feels like it failed. Compare it to a post with 200,000 views and the instinct is obvious: the big number worked, the small one did not. That instinct is measuring the wrong thing.
Google's AI Overviews cut the average click-through rate on informational search results by 61 percent, according to a study of 3,119 queries across 42 organizations. That number sounds like a death sentence for anyone making how-to content. Buried in the same report is a different number: brands cited by name inside the AI Overview saw their click-through rate rise 35 percent, not fall. The traffic did not disappear evenly. It disappeared from everyone except the source Google actually named.
Searching how to switch tabs with keyboard shows exactly what that looks like in practice.
The answer itself is generated by Google. The citation next to it, and the panel beside it, both point to one YouTube channel. Nobody ranked for this. Somebody got named.
The same pattern held on a completely different kind of query.
How to validate idea before building pulled up the same structure: a named video citation inside the answer, and the same channel again in the sidebar, next to a newsletter and a couple of articles.
- Both queries cited a specific YouTube video by name inside the AI Overview, not a generic web link
- Both queries also surfaced a separate, dedicated video panel next to the answer
- Neither citation went to the channel with the most subscribers, it went to whichever video actually answered the specific question
The 61 percent decline is also already out of date. A follow up covering 5.47 million queries through February 2026 found organic click-through on AI Overview pages climbing back, up 85 percent from its low point two months earlier. The panic phase of AI search is ending faster than the trend it was reacting to.
Rand Fishkin, who has tracked zero click search for years, put the actual risk plainly: if your website and business model are things Google and AI can disintermediate, they will. The businesses being disintermediated are the ones that had nothing distinct enough to be named. Being the cited source is the opposite of being disintermediated. It is Google doing your distribution for you, on the one query that matters most: the exact problem your product solves.
This matters more for some products than others. Consumer research has a long-standing distinction between low involvement purchases, cheap and easy to judge on sight, and high involvement purchases, the kind people research before committing to. A cool t-shirt sells fine from a five second ad. A piece of software, or anything with real switching cost, usually does not. A large field experiment across 71 products and 25 categories found exactly this split: clothing, cars, and food converted well from social ads, while financial services, electronics, and apps did not. A tutorial is not a workaround for that gap. It is the thing that closes it, because it lets someone see the product work before they have to trust a claim about it.
200 views is not a small number if it is the 200 people who typed the exact sentence describing your product's problem, and Google decided your video was the answer worth naming. The view count was never the real scoreboard. Being the citation was.