Answer Engine Optimization Has to Beat 0.57 Clicks Per Search

The open web now receives 0.57 clicks for every search performed. Not per user. Per search. Across a basket of twelve common US queries, eight of the twelve return fewer than one click each. That is the number answer engine optimization is actually working against, and almost no definition of it says so.

The definitions on offer are all versions of "optimise to be the answer, not a result". True, but it makes the work sound like an upgrade to SEO rather than a response to something specific. So let us start with the size of the hole.

Eight of twelve queries return fewer than one click

The basket is twelve US queries covering a range of intents, measured with clickstream data – meaning real observed clicks, not estimates.

Eight return fewer than one click per search. The worst turns a large monthly search volume into a small number of clicks. The best clears one, for an unglamorous reason: the searcher opens several results to compare them, which only happens where comparing is the task.

Split the basket by intent and the picture stops being a general decline:

A factor of 3.1 between them.

The click did not disappear from search. It disappeared from the part of search where the answer was short enough to print on the page.

That is the real brief, and it is far more useful than "be the answer".

One row that should change how you read your reports

python list comprehension returns one of the healthiest clicks-per-search figures in the entire basket. On separate trend data, its volume is down 62.3% from its peak year.

Both are true, and together they are a trap.

The casual majority stopped searching it – they ask an assistant now. What remains is the residue: the smaller group who wanted the documentation page all along, and who still click.

A healthy click-through rate on a collapsing query is a warning, not a win. If your dashboard shows click-through holding steady while sessions fall, you are probably looking at exactly this and calling it resilience.

What answer engine optimization actually has to do

Three things, in order.

1. Be a source the answer is built from. Not near it. In it. There is no position two inside a paragraph, and the runner-up gets nothing.

2. Be quoted repeatedly, not once. Take the pages of a site that get quoted at all, then count how many times each one is quoted. Wikipedia averages 8.11 citations per page. YouTube averages 1.87.

In other words, assistants keep coming back to the same Wikipedia page for question after question, while a YouTube page tends to be used once and never again. Dense, well-organised pages get returned to. Sprawling ones do not.

That ratio is the clearest format instruction available. Write so a single passage can be lifted out and still be correct on its own. The unit of work is the passage, not the page.

3. Concentrate where a click still exists. If a query returns 0.39 clicks per search, winning it outright still wins you very little. The commercial end of the basket returns three times more.

Two adjacent findings sharpen point two. In a software-category probe, 43 comparison sites nobody tracks out-cited G2, Capterra and TrustRadius combined – small sites with no authority, winning on format. And in categories that publish their prices, the vendors themselves take 90.9% of the citations, against 48.7% where pricing is hidden behind a sales call.

The mechanic is the same in both. What gets quoted is what can be checked. A page with a price on it can be quoted with confidence. A page that says "contact us for pricing" cannot be quoted at all.

The uncomfortable arithmetic on citations

The standard pitch is: you lose the click, you gain the citation, the citation builds the brand.

Test it. In browsing data across 68,879 searches, clicks on the sources inside an AI summary ran at 1%.

One per cent is not a channel. If you are pricing this work on the expectation that citations turn into visits, you are pricing on a rate nobody has demonstrated. Price it as brand exposure – genuinely valuable – and the business case survives contact with the data. Why citation share is not traffic explains why the platform citing you most is frequently not the one sending you anyone.

There is a real counterweight. AI referral traffic exists, it is growing at 117.4% a year, and it converts about 54% better than non-AI sources for US retail. It is also a low single-digit share of most sites' visits, because only 6.8% of assistant conversations include a web citation at all. Model it as a new channel with its own economics: about a fiftieth of the volume, better conversion, growing fast.

A measurement problem most tools have not caught up with

Since late August 2026, Google has routed result clicks through an intermediate redirect that blocks results-page scraping. Rankings are unchanged. What third-party tools can see is not.

Clickstream panels – the source of the 0.57 figure – are unaffected, because they watch the browser rather than scrape the page. Rank trackers are affected.

Your most reliable first-party record is Google's own. The Search Console Performance report shows impressions, clicks and average position for your site from Google's side of the transaction, and traffic from AI features is included in the standard "Web" search type rather than broken out separately. It is not a complete picture of AI visibility, but it is the one dataset nobody can scrape their way around.

A 90-day answer engine optimization plan

Answer engine optimization does not need new software to start. It needs one analyst and one quarter.

Weeks 1–2. Measure the hole. Pull clicks per search for your top 200 queries – clicks divided by volume, not impressions or average position. Rank them. Almost nobody has done this.

Weeks 3–4. Split the list. One question: can a model finish this user's task without them leaving the tool they are in? Yes goes in the forecast-a-decline pile. No goes in the invest pile.

Weeks 5–8. Rebuild the top twenty. Make them canonical, comparison-shaped and checkable. Real figures, dated. Tables that can be quoted. Publish the numbers you currently gate, because gated facts are citations handed to someone else.

Weeks 9–12. Baseline across assistants. Measure citations on more than one, and report the disagreement rather than averaging it. The six assistants probed disagree sharply: on one site, the share of its citations coming from a single assistant ran from 1.2% to 54.2%.

Questions people ask about answer engine optimization

These are the questions searched most often alongside answer engine optimization, answered from the click and citation data above.

What is answer engine optimization? It is the work of remaining a source when the results page has stopped being an index and become a destination. In practice: writing pages that an AI answer can quote confidently, on the queries where a click still exists.

How does answer engine optimization work? Answer engines retrieve pages that are already indexed, then lift passages from them into a generated answer. So the work is making individual passages independently correct, checkable and quotable, rather than making a whole page rank.

What is the difference between AEO and SEO? SEO optimises for a crawler that ranks pages against a query. AEO optimises for a retrieval layer that quotes passages into an answer. The unit changes from the page to the passage, and there is no second place.

Does AEO replace SEO? No. Pages still have to be crawlable, indexed and snippet-eligible before they can be quoted, so the fundamentals still apply. AEO is what you add on top, and it matters most on the queries where clicks have collapsed.

Is answer engine optimization worth the investment? It depends entirely on how you value the outcome. Citations convert to visits at about 1%, so if you fund this as a traffic channel it will disappoint. Fund it as brand presence in the place buyers now form shortlists, and the case holds.

---

Clicks-per-search figures: Ahrefs clickstream data across a 12-query US basket, pulled live 7 September 2026, with clicks per search recomputed from raw clicks divided by raw volume. Citation figures: live Ahrefs index pull across six answer platforms and five domains. Click-through inside summaries: Pew Research Center, 68,879 searches. Full method in the report, How AI Changed the Way People Search.