When clicks fall: how to judge link building and SEO in a zero-click world

If your search clicks are falling, how do you know whether your SEO and link building are still working? This guide separates the numbers that have become unreliable from the ones that still show progress, and gives you a monthly report you can copy.

What the research says about fewer clicks

Start with the best independent evidence we could find, and be clear about what it covers.

Pew Research Center, July 2025. Pew tracked the browsing of 900 U.S. adults who belong to an online panel and agreed to install a tracking app. It looked at 68,879 unique Google searches from March 2025. About 18% of those searches (12,593) produced an AI summary. Pew found that people who saw an AI summary clicked a traditional result link in 8% of visits. People who didn't see one clicked a result in 15% of visits. Clicks on a link inside the AI summary itself happened in 1% of visits. And the browsing session ended after the page on 26% of pages with an AI summary, against 16% of pages with only traditional results. Pew counts a session as ended when the person exits the browser for five seconds or more.

Clicks and exits, with and without an AI summary

• Clicked a regular result, no AI summary15% of visits
• Clicked a regular result, AI summary shown8% of visits
• Ended the session, no AI summary16% of pages
• Ended the session, AI summary shown26% of pages

What this doesn't show: it covers Google only, U.S. adults only, and one month in 2025. It says nothing about any single website, and it can't tell you whether the person who didn't click got what they needed.

Ahrefs, February 2026. Ahrefs compared 150,000 keywords that triggered an AI Overview in December 2025 with 150,000 informational keywords that didn't, using aggregated Search Console data (average desktop click-through rate, which is clicks divided by impressions). For the page ranking first, the rate on keywords with an AI Overview fell from 0.073 in December 2023 to 0.016 in December 2025. On keywords without one it fell from 0.076 to 0.039. Ahrefs' headline figure is 58% lower, up from 34.5% in its April 2025 version of the study. Limits: desktop only, only keywords that had an AI Overview in December 2025, and Ahrefs sells SEO software.

Google, August 2025. Liz Reid, who leads Google Search, wrote that total organic click volume from Google Search to websites "has been relatively stable year-over-year" and that average click quality has increased. The post gives no figures, charts or method for that statement.

So the sources disagree. Pew and Ahrefs measure click behavior on results pages that show AI summaries. Google makes a claim about total clicks across all of Search, without showing its work. Both can be partly true: individual results can lose clicks while the total holds up. Nobody outside Google can check the total.

Our view: you don't need to settle that argument. You need a measurement setup that works if clicks keep falling and still works if they don't.

Why a falling click count doesn't mean your SEO failed

A click needs three things in a row: your page is shown, the person chooses it, and they leave the results page for yours. AI summaries mostly interfere with the second and third steps. Your page can be shown just as often as before and still get fewer visits.

Other things also move clicks, with no AI involved: seasonality, a rival's new page, a change in how Google lays out the page, and your own edits. Pew's finding that people exit more often after an AI summary is a reminder that "no click" can mean "question answered", not "page ignored".

That's why a single number, such as organic sessions, can't tell you whether your SEO and link building work. It tells you how many people arrived. It doesn't tell you how many people saw you, remembered you, or came back later by typing your name.

What Google says about AI features in Search Console

Many people assume AI traffic is hidden. It's more accurate to say it's blended in. Here is what Google's own pages say.

• AI features are inside the normal report. Google's page on AI features (updated December 10, 2025) says sites that appear in AI features such as AI Overviews and AI Mode "are included in the overall search traffic in Search Console" and are reported in the Performance report within the "Web" search type.
• Clicks and impressions follow the usual rules. Google's Search Console help says a click on a link to an external page in an AI Overview or in AI Mode counts as a click, and that standard impression rules apply.
• Position is odd. For AI Overviews, Google says an AI Overview "occupies a single position in search results, and all links in the AI Overview are assigned that same position." For AI Mode, position follows the same method as a normal results page.
• There is a separate, narrower report. Google added a generative AI performance report. Its help page says it shows how many times links to your site were shown to a user in a generative AI feature on Google Search, and that, as of August 31, 2026, Google has rolled it out to all websites worldwide, though not every site may see it yet. It includes data from the Web search type, so it overlaps with the main report. The page describes impressions, by page, country, date and device. It doesn't describe clicks, and the usual 1,000-row limit applies. The newest data can be preliminary.

Two practical points follow. First, you can't pull "AI clicks" out of the main Performance report; they sit with everything else in Web. Second, the AI report counts shown links, not visits, so a rise there is a sign of being included, not a sign of traffic.

Numbers that still show progress

None of these proves that a link caused a result. Together they tell a story that a click count alone can't.

New referring domains, and who they are

Count the sites that newly link to you each month, and note whether they're relevant to your topic and market. This is the most direct measure of the work you paid for or did. Our guide to link quality metrics explains how to judge a site without leaning on one score, and how to measure link building success covers checking each placement.

Where a link was paid for, keep in mind that Google's spam policies treat links that are bought or sold to pass ranking signals as link spam, unless the link is marked rel="sponsored" or rel="nofollow". Some placements carry a label. Record which ones do, so your count is honest.

Referral visits from the sites that link to you

A link that sends no one is still a link, but one that sends real readers is a link people actually saw. In your analytics, look at sessions from referral sources and see which linking pages bring readers who do something useful. Small numbers are normal. Judge direction over several months.

Branded searches

When more people search your name, your products or your domain, someone has heard of you. Search Console has a branded queries filter in the Performance report. Google's help defines branded queries as those that include your brand name, domain, or a brand-specific product or service, including common misspellings and variations. The filter isn't available for sub-properties (such as a site's /blog/ section) or for sites with few impressions, and its data starts from March 11, 2025. Compare branded impressions and clicks across quarters, not weeks.

This number is the closest cheap proxy for "more people know us", and it holds up when AI summaries answer the general questions.

Impressions versus clicks, side by side

Put impressions and clicks for the same pages next to each other and watch the gap. Rising impressions with flat or falling clicks is the pattern you'd expect when summaries take some of the clicks. Falling impressions is a different problem: you're being shown less. Don't judge by average position alone, for the reason above.

Calls, form enquiries and assisted conversions

Count the enquiries that matter to the business, and ask new customers how they found you. A required "How did you hear about us?" field on a contact form costs nothing. In Google Analytics 4, the attribution reports let you compare attribution models side by side, which helps you see organic search or referral visits that came earlier in a path and didn't get the last click.

What to watch, and what to stop leaning on

Still shows progress

• New relevant referring domains
• Referral visits from linking pages
• Branded searches
• Impressions next to clicks
• Enquiries and assisted conversions

Don't rely on alone

• Organic clicks as the only score
• Average position
• One keyword's rank
• DR or any single score

What to stop relying on

• Organic clicks as the only score. They still matter, because visits can turn into customers. But they now mix two things: how well you rank and how often a summary answers the question first.
• Average position. Because all links in an AI Overview share one position, an average across many searches can shift for reasons that have nothing to do with your page.
• A rank for a single keyword. Look at a group of related terms.
• Any one third-party score. DR, DA and similar scores are estimates from tool makers, not Google measures.
• "AI visibility" scores as proof. Answers from AI tools vary from one run to the next. Our guide to checking whether AI tools mention your brand explains why one answer proves little.

A worked example with made-up numbers

These numbers are invented to show the reasoning. They aren't from a real client.

A small software company reviews April against January:

• Impressions in Search Console: 100,000 in January, 120,000 in April.
• Clicks: 3,000 in January, 2,700 in April.
• Click-through rate: 3.0% in January (3,000 ÷ 100,000), 2.25% in April (2,700 ÷ 120,000).
• Branded impressions: up 15%.
• New referring domains: 14 relevant sites over the quarter, 3 of them sending a few visits.
• Demo requests: 40 in January, 44 in April.

A click-only view says "traffic fell 10%, SEO isn't working." The wider view says the site was shown 20% more often, a smaller share of people clicked, more people searched the brand name, the link profile grew with relevant sites, and enquiries rose 10%. That isn't proof that links caused the rise in enquiries. It is enough to say the program isn't failing, and enough to keep going and look again next quarter.

The reverse case matters too. If impressions, branded searches and enquiries all fall while clicks fall, there is a real problem to investigate.

A monthly report template you can copy

Keep it to one page. The same fields every month matter more than clever charts.

How to read the report without fooling yourself

• Compare with the same month last year. Seasonality can look like success or failure.
• Keep a change log. If you rewrote a key page in March, say so. Otherwise you'll credit or blame the wrong thing.
• Look at quarters for links and brand searches. They move slowly.
• Wait before drawing a conclusion from one month. The Pew study is a snapshot of March 2025, and Google changes how it shows results.
• Write down what you can't tell. For example: "We can't separate AI Overview clicks from other Web clicks."

1. 1Pull the numbersSearch Console, analytics and enquiries
2. 2CompareLast month and the same month last year
3. 3Check the change logDid we change anything?
4. 4Write one sentenceWhat the data does and doesn't support
5. 5Pick one actionDecide what to do next month

What to do

1. Open Search Console and note impressions, clicks and click-through rate for your ten most important pages. This is your starting point.
2. Turn on the branded queries filter (if your site has it) and record branded impressions and clicks.
3. Check whether the generative AI performance report appears for your property.
4. List your referring domains from the last three months and mark the relevant ones.
5. Add "How did you hear about us?" to your contact form, and record calls and enquiries in one sheet.
6. Copy the template above into a document and fill it in for last month.
7. Stop reporting organic clicks as the only measure of success.

If you want help building the links you're tracking, you can read about our services, where you choose the sites and approve the content before anything is published. For the bigger picture of where search is heading, see our posts on whether AI search replaces SEO and how AI assistants pick their sources.

Sources

• Pew Research Center: Google users are less likely to click on links when an AI summary appears in the results (July 22, 2025; 900 U.S. adults, 68,879 Google searches, March 2025)
• Ahrefs: AI Overviews reduce clicks by 58% (update) (February 4, 2026; 300,000 keywords, December 2023 and December 2025)
• Google: AI in Search is driving more queries and higher quality clicks (August 6, 2025)
• Google Search Central: AI features and your website (updated December 10, 2025)
• Google Search Console Help: Clicks, impressions and position (AI Overviews and AI Mode sections)
• Google Search Console Help: Generative AI performance report (Search) (rollout to all websites as of August 31, 2026)
• Google Search Console Help: Performance report, common tasks and use cases (branded queries filter)
• Google Spam policies for Google web search (updated August 28, 2026)
• Google Analytics Help: Advertising and attribution (attribution models report)

Checked on 6 October 2026.