Everyone wants the same number: if my page is cited in an AI Overview, how many people actually click through. The honest answer is uncomfortable, very few do, and the best public data puts clicks on links inside the summary at around one in a hundred visits. That does not make AI Overview citations worthless, but it does mean you should value them for the right reason. Here is what the data actually shows about clicks on AI Overview citations, why the number is so low, and what it means for how you measure success.
The short answer
Very few people click on AI Overview citations. The most cited public study, from Pew Research, found that users clicked a link inside the AI summary in only about 1 percent of visits to pages with such a summary, and that overall they clicked any search result far less often when an AI summary appeared. So if you are hoping AI Overview citations drive large referral traffic, the data says expect little. Their value is mostly visibility and authority, being seen as the cited answer, rather than clicks. There is no single universal absolute number of monthly clicks, since it depends on your traffic, but the click rate on citations is low, and you should plan around that.
What the data actually shows
Start with the numbers, because guesses abound. Pew Research, analyzing real browsing behaviour, reported that users clicked a traditional search result in 8 percent of visits when an AI summary was present, versus 15 percent when it was not, and clicked a link inside the AI summary itself in just 1 percent of visits. Those are the figures to anchor on: roughly half the overall click-through when an AI summary appears, and a very small rate of clicks on the summary’s own citations. So the realistic expectation for AI Overview citation clicks is low, single-digit percent at best for the surface and around 1 percent for the in-summary links.
The click rates at a glance
The contrast is stark in a table.
| Scenario | Click rate (Pew) |
|---|---|
| No AI summary, a search result clicked | 15% |
| AI summary present, a search result clicked | 8% |
| A link inside the AI summary clicked | about 1% |
Read it plainly: when the AI summary appears, overall clicks roughly halve, and the citations within it are clicked rarely. This is the empirical basis for treating AI Overview presence as a visibility play more than a traffic play.
Why so few people click
The low rate is not a glitch; it is the design working as intended. An AI Overview exists to answer the question on the spot, so when it succeeds, the user has what they need and has no reason to click. Google itself frames AI features as resolving queries more thoroughly, and a page only needs to be indexed and eligible for a snippet to be used in them. So the better the AI answer, the less the click, which is the opposite of the classic search bargain where the link was the destination. Understanding that the answer is now the destination explains why citation clicks are low and reframes what you should expect from being cited.
The scale context
A low rate is not the same as zero impact, because the surface is enormous. Google reported that AI Overviews reached 2 billion monthly users, so even a small click rate applies across a vast base, and being the cited source means exposure to a large audience whether or not they click. The honest framing is that the percentage is low but the reach is high, so the value shows up as visibility and brand presence at scale rather than as a flood of referral clicks. Do not multiply the rate by the whole base and claim a precise click count, since your slice depends on your queries, but do recognize that citation exposure at this scale has real value.
What this means for strategy
The data should change how you measure, not whether you participate. If you judge AI Overview citations by referral clicks alone, you will undervalue them and may wrongly conclude AI search does not matter. Instead, treat citations as a visibility and authority metric: being the cited answer builds awareness and trust at the moment of the question, which influences later branded searches and direct visits. This is the same reasoning behind treating AI presence as its own scorecard, covered in classic SEO versus AI visibility. Plan for low direct clicks, and measure the value of being the answer, which is the real return AI Overview citations offer.
Does being cited still have value
Yes, clearly, just not mainly as clicks. Being cited puts your brand in front of the user as the trusted answer, which builds awareness and credibility even without a click, and it positions you ahead of competitors who are absent. There is also a compounding effect: brands repeatedly seen as the cited answer become the names users later search for directly. So the value of an AI Overview citation is real but indirect, brand exposure, authority, and downstream demand, rather than immediate referral traffic. Judged on that basis, citations are worth pursuing; judged only on clicks, they will disappoint. The case for pursuing them is in why GEO is important.
How to think about ROI now
Because the clicks are few, your measurement has to mature. Treat AI Overview citation as a leading indicator of visibility, tracked as citation presence and share of voice, and correlate it over time with downstream signals like branded search volume, direct traffic, and pipeline, rather than expecting a clean click-to-conversion line. This is the same leading-indicator logic discussed in are ChatGPT citations worth tracking. The mistake is applying a last-click traffic model to a surface designed to answer without a click. Measure presence and its downstream correlation, and the ROI of AI Overview citations becomes legible even though the direct clicks are sparse.
When people do click
It is worth knowing the exceptions, because not all queries behave the same. Users are more likely to click through when the AI answer is incomplete, when the question is complex or high-stakes and they want to verify, or when they need something the summary cannot provide, like a tool, a purchase, or detailed documentation. So content that clearly offers more than the summary, depth, specifics, or an action, can still earn the clicks that do happen. The takeaway is not to abandon click-worthy content but to recognize that the clicks concentrate where the summary is insufficient, and to make your page the obvious next step when a user does decide to go deeper.
Why ranking does not save the click
Some assume a top ranking restores the lost clicks, but the data says otherwise. Citations and clicks both spread beyond the top of the rankings, and Ahrefs found that only 38 percent of AI Overview citations come from top-10 pages, down from about 76 percent a year earlier, while the click suppression applies regardless of position. So ranking first does not exempt you from the low citation-click reality, and being cited from a lower position is increasingly possible. The implication is to pursue citation broadly, since position alone neither guarantees the citation nor restores the click. The ranking nuance is in do you need to be in the top 10 to rank in AI Overviews.
A worked example
A team measured AI Overview citations purely by referral clicks and nearly cut their AI search effort, since the clicks were tiny. Before doing so, they reframed the metric: they tracked how often they were the cited answer across their key questions, watched their branded search and direct traffic over the following months, and found both rising as their citation presence grew. The citations were not driving many direct clicks, but they were building awareness that showed up downstream. Had they judged on clicks alone, they would have abandoned a channel that was quietly working. The low click rate had been real; the conclusion to quit had been wrong. The deciding move was simply changing the yardstick: once they measured presence and its downstream effect instead of raw citation clicks, the channel’s contribution became visible, and the team kept investing in the content and authority that grew it.
Common misconceptions
A few myths distort the picture. The first is that a citation should drive lots of referral traffic, when the data shows around 1 percent click the in-summary links. The second is that low clicks mean citations are worthless, when their value is visibility and authority. The third is that a top ranking restores the clicks, when suppression applies regardless of position. The fourth is that you can multiply the rate by total users for a precise click count, when your slice depends on your queries. The fifth is using a last-click model on a no-click surface. Clear these and you measure AI Overview citations sensibly, and you stop walking away from a channel that pays off in visibility rather than raw clicks.
The bottom line
Very few people click on AI Overview citations, with the best public data putting clicks on in-summary links at around 1 percent of visits and overall result clicks roughly halving when a summary appears. That is by design, since the AI answer is now the destination. But with billions of monthly users, being the cited answer delivers real visibility, authority, and downstream demand even without the click. So plan for low direct clicks, measure citation presence as a leading indicator correlated with branded search and pipeline, and value AI Overview citations for the exposure they give rather than the traffic they do not.
Frequently asked questions
How many people actually click on AI Overview citations?
Very few. The most cited public study, from Pew Research, found that users clicked a link inside the AI summary in only about 1 percent of visits to pages with such a summary, and that overall they clicked a search result in 8 percent of visits with an AI summary versus 15 percent without. There is no single universal absolute number of clicks, since it depends on your traffic, but the rate of clicks on AI Overview citations is low, so plan accordingly.
If clicks are so low, are AI Overview citations worth it?
Yes, but for visibility and authority rather than referral traffic. Being the cited answer puts your brand in front of a vast audience at the moment of the question, building awareness and credibility and positioning you ahead of absent competitors, which compounds into later branded searches and direct visits. So citations are worth pursuing when judged as a visibility and brand metric. They disappoint only when judged purely on immediate clicks, which the surface is designed to minimize.
Why do so few people click when an AI summary appears?
Because the AI summary is designed to answer the question on the spot, so when it succeeds the user has what they need and no reason to click. That is the opposite of classic search, where the link was the destination. The better the AI answer, the fewer the clicks. Users still click when the summary is incomplete, the question is high-stakes and they want to verify, or they need something the summary cannot provide, like a tool, purchase, or detailed documentation.
How should I measure success if not by clicks?
Treat AI Overview citation as a leading indicator of visibility: track how often you are the cited answer and your share of voice across key questions, then correlate that over time with downstream signals like branded search volume, direct traffic, and pipeline. Avoid applying a last-click, traffic-only model to a surface built to answer without a click. Measuring citation presence and its downstream correlation gives you a legible ROI even though the direct clicks on citations are sparse.
