~800K impressions, almost no clicks… but 15-20 signups/day. What’s actually working.
Launched this on March 19.
For the first ~4 weeks, nothing really happened.
No signups. No signal. Just publishing and waiting.
Then things started to move.
Now it’s around 15–20 signups/day, and ~70–75% of them actually leave reviews.
That’s been surprisingly useful for AI tool founders — they start getting real user feedback almost immediately after listing.
What’s confusing at first:
I’m getting hundreds of thousands of impressions on Google, but almost no clicks.
Interestingly, most signups are coming from ChatGPT and Perplexity citing the content
Google is still sending impressions.
But the actual users are coming from AI search — ChatGPT, Perplexity, etc.
So I stopped optimizing only for “ranking on Google”
and started focusing more on being useful inside AI answers.
Right now, I’d say ~80% of my effort goes there.
What’s interesting is:
when you structure content for clear answers and decisions,
it still ends up working on Google anyway.
The shift for me was simple:
Less “content for traffic”
More “content that helps someone decide”
Things like:
-comparison pages
-direct answers upfront
-clarity over volume
Still early, but one thing feels clear:
It’s not just about being found anymore.
It’s about being used — whether that’s on Google or inside an AI answer.
Curious if others are seeing this shift too.
This matches what I have started noticing as well (at a small scale though).
The shift from “content that ranks” to “content that gets used” feels real to me somehow, especially when the goal is helping someone make a decision, not just bringing them to a page.
I am working on a blog in a different space, and I have been trying to focus more on clarity and direct answers rather than volume. Things like explaining a specific problem clearly and showing what actually breaks in practice.
What is interesting is that this kind of content seems to work across both directions. It is easier to surface in AI answers, and it still performs fine on search.
Curious, are you structuring your content in any specific way for AI answers, or is it mostly about writing clearly and letting it surface naturally?
Honestly, I don’t think there’s a secret AI-optimization format yet.
Most of the lift seems to come from making pages extremely easy to extract answers from:
One thing I noticed is that AI systems seem to prefer pages where the intent resolution is obvious.
If someone asks “which tool is better for X,” the page that directly resolves that question tends to surface more often than long-form content trying to rank broadly.
Still experimenting though — feels like everyone is learning this in real time.
The interesting shift is that “SEO” is starting to split into two jobs:
Google decides who gets seen.
AI decides who gets used.
A page can underperform on clicks and still win if it becomes the source models keep pulling into answers.
That changes the optimization target.
You’re no longer just writing to rank.
You’re writing to become the answer layer.
And that usually rewards the same things:
clear comparisons, direct conclusions, less fluff, stronger structure.
Traffic used to be the goal.
Now citation is.
Yeah and I think that’s the part many people still underestimate.
A lot of pages that look weak from a traditional SEO perspective are still influencing decisions because they’re getting surfaced inside AI answers.
Which means the value is no longer only in the click.
Sometimes it’s in becoming the source that shapes the recommendation itself.
Feels like visibility and influence are starting to diverge a bit for the first time.
And that split usually becomes the ceiling.
If the asset is really “the source models keep citing,” then the real question is whether this stays a content brand or becomes infrastructure.
Because those compound very differently.
One gets read.
The other becomes part of how decisions get made.