AI Search Is Cutting Clicks. Small SaaS Teams Should Measure Signups Per Page.
Search demand did not disappear. The click did.
That is the part a lot of small SaaS teams are reading wrong right now. If AI answers handle the casual research on the results page, then your content should not be judged by how many tourists it attracts. It should be judged by whether it helps someone sign up.
How should SaaS teams measure content in AI search?
Measure content by signups, demos, and assisted conversions per page, not raw sessions. Umesh Malik argues that AI search changed the click pattern more than the underlying demand. His source roundup says zero-click searches reached roughly 68% in early 2026, and Google AI Mode showed a ~93% zero-click rate.
That matches the shift a lot of founders are feeling. Informational queries still happen. People still ask the questions. They just get enough of the answer without leaving the interface.
If your dashboard still treats pageviews as the main win, you will undervalue the pages that help real buyers move. A page with 200 visits and 10 demos is worth more than a page with 5,000 visits and no pipeline. That sounds obvious. A lot of teams still report the opposite because sessions are easy to screenshot.
The clicks that survive are often buyer clicks
AI referral traffic is small, but the people who do click through look much closer to buyers than browsers. Umesh Malik cites Similarweb clickstream data from Apr–May 2026 showing ChatGPT referral conversion at 7.1%, second only to paid search. He also cites Ahrefs data showing AI search drove 0.5% of traffic but 12.1% of signups.
That changes the way a tiny SaaS team should read declining traffic. Fewer visits can still mean better content economics if those visits come from people comparing tools, checking proof, or validating a purchase.
The practical mistake is treating every lost click like lost opportunity. A big chunk of those lost clicks were never going to buy. AI answers are eating broad, early research. What gets through is often the person who still needs specifics your site can provide:
- proof that your product works
- pricing details
- migration steps
- comparison context
- limits, tradeoffs, or implementation detail
Those are not vanity visits. Those are sales-adjacent visits.
Bottom-funnel pages now carry more of the content ROI
Pages built for decision-stage questions are better aligned with how AI search filters traffic. Umesh Malik makes the case that winning AI search is about being cited inside answers and then earning the occasional click that follows. That means the page has two jobs now: feed the answer engine clear, quotable facts, and help the human finish the decision once they arrive.
For a small team, that usually means shifting effort toward assets like:
- comparison pages
- use-case pages
- migration guides
- pricing explainers
- founder-written case studies with real numbers
The old traffic play was often top-of-funnel volume. The newer play is trust density. Give the model facts it can extract. Give the buyer evidence they can verify.
Malik also points to peer-reviewed GEO research showing that quotations lifted AI visibility 41%, statistics 32%, cited sources 30%, and fluency 28%. Those are useful inputs for writing pages that can both get cited and convert once clicked.
Last-click reporting will make you kill the pages doing real work
A lot of content now helps before the visit or between visits, which makes last-click attribution worse than usual. Umesh Malik warns that teams need to separate AI search in reporting or they will reach the wrong conclusion about performance.
If someone sees your brand in an AI answer, comes back later direct, and then starts a trial, the page that shaped the decision may never get credit in a basic SEO report. The same goes for a comparison page that gets skimmed once and revisited later during procurement.
Small teams do not need a fancy attribution stack to handle this better. They do need to stop asking only one question: “How much traffic did this post get?” Better questions are:
- How many signups or demos came from this page?
- How often does this page show up on paths that lead to signups?
- Do visitors from this page convert at a higher rate than site average?
- Does this page answer a decision-stage question better than competitors do?
If AI keeps broad research inside the answer box, then raw traffic becomes a weaker success metric. Signups per page gets closer to the truth.
That is the shift. Do not panic because clicks are down. Tighten the pages that help buyers choose, and measure them like revenue assets instead of blog posts.