If your branded clicks have fallen while rankings have barely moved, the likeliest explanation since 26 September 2026 is an AI Overview now appearing when people search your own name. Search Console will not label it for you. You confirm it by isolating branded queries, comparing CTR before and after that date, checking the results page by hand, and ruling out the duller causes first.

Below is how we diagnose it for clients, what the headline data does and does not prove, and how to report it without sending a board into a spin.

Branded AI Overviews jumped in late September, but that is presence data, not click data

On 1 October 2026 DemandSphere published "AI Overviews on branded queries tripled in September". AI Overview presence on its tracked branded keywords averaged 26.9% across 1 to 5 September, then hit 82.1% on 29 September. The highest single day was 90.5% on 27 September.

The timing matters for your diagnosis. DemandSphere saw isolated spikes on 18, 20, 23 and 24 September, and each fell back to a baseline of roughly 30% within a day. The sustained rise began on 26 September. Search Engine Land reported the finding the day DemandSphere published and noted that Google had announced no change. Search Engine Roundtable separately saw Overviews return for many large brand names.

So the shift is real and well corroborated. What it does to your clicks is a separate question.

What DemandSphere measured, and what it did not

DemandSphere counted whether an Overview appeared. It did not count clicks. An Overview was logged whether or not it cited the brand being searched, so one that sent people straight to your site counts the same as one that sent them elsewhere.

The other limits are worth stating plainly:

  • Keywords were tracked once a day, with all markets and devices combined.
  • There is no breakdown by country, device or industry.
  • DemandSphere has not disclosed how many keywords it tracks.

Above all, the tracked set is DemandSphere's, not yours. Your brand may not be affected at all. A smaller brand, a different market or a query with strong sitelinks could behave quite differently. In the coverage we have seen, nobody had published branded click data from this change.

An Overview appearing on your name is a fact about the results page. Whether it costs you anything is a fact about your own data.

Why the Ahrefs and Pew averages do not transfer to your brand

Two figures get quoted constantly. Ahrefs' February 2026 update, based on December 2025 data across 300,000 keywords, found that AI Overviews correlate with a 58% lower CTR for the top-ranking page. Pew Research, studying 68,879 searches in July 2025, found that users clicked a result in 8% of visits with an AI summary against 15% without, and only 1% clicked a link inside the summary.

Both are useful. Neither is about brands. In the Ahrefs study, 99.2% of the keywords showing an Overview were informational. Someone asking how to descale a kettle can happily stop at a summary. Someone typing your company name usually wants your site, your login page or your phone number. We think these averages probably overstate the effect on branded navigational searches. Do not apply 58% to your brand and present it as a forecast.

What Search Console can and cannot tell you

Search Console is still the best evidence you have, as long as you read it as circumstantial rather than conclusive.

The impressions-up, clicks-down signature

The classic AI Overview pattern is impressions flat or rising while clicks and CTR fall, with average position roughly unchanged. A ranking loss looks different. Impressions and position fall together, because you are being shown less and lower.

One vendor blog suggests that because Overviews cite several sources at one shared position, more pages may now earn an impression per query than before. That is a single source, so treat it as a possibility, not an explanation. It is still one reason impressions can rise while clicks fall.

Check your own property for AI reporting before you rely on it

This is where advice online gets muddled. One source says AI Overview and AI Mode traffic is counted inside the normal Web search type, with no separate filter. Another claims Search Console now shows AI Overview and AI Mode queries with their own impressions, clicks and CTR, while admitting the relevant Search Appearance filter is not universally available. A third mentions a "generative AI features" report.

Do not promise a report you have not seen

We could not verify any dedicated AI Overview report against Google's own documentation. Open your property and look. If a generative AI or AI Overview view exists for you, use it. If it does not, assume Overview traffic is blended into Web search and infer it from the pattern described below.

A step-by-step diagnosis for branded click loss

This is the sequence we run. It takes an afternoon if your property is in reasonable shape, and it is the same process we use in our SEO services when a client flags a sudden branded drop.

Split branded from non-branded, then before and after 26 September 2026

In the Performance report, add a query filter using a custom regex that matches your brand name, its common misspellings and any product names people search alongside it. Then build the inverse filter for non-branded queries. For example:

text
(?i)(devonic|devonic web|devonicweb|devonik)

Swap in your own terms, then check the matching queries list so you are not catching unrelated words.

Next, use compare mode. Set the period after 26 September 2026 against an equal-length window before it, and repeat against the same period last year if seasonality affects you. Keep the query set stable, so new long-tail queries do not skew the averages. Record clicks, impressions, CTR and position for the branded segment.

If branded CTR has fallen and non-branded has not, something is specific to brand results. If both have fallen together, look wider.

Segment by device and band CTR by position

Add a device comparison. If desktop CTR collapses while mobile holds, or the reverse, the format of the results page is a factor, not a general loss of interest.

Then stop looking at one site-wide CTR. Export branded queries and group them by position band: position 1, positions 2 to 3, and 4 to 10, say. A branded query at position 1 with a CTR well below its own previous level is your strongest signal. Compare each query with its own history, not a generic threshold from a blog post, because branded CTR varies enormously between brands.

Check the SERP by hand

Data shows a pattern. Only the results page shows the cause. Search your top branded queries in a private window, on desktop and mobile, and from the markets you serve if you can. Note whether an Overview appears, what it says, and whether it links to you.

Screenshot and date everything. To make this repeatable rather than a one-off, our guide to tracking your visibility in Google AI Overviews covers monitoring over time.

Rule out the other causes first

The AI Overview story is plausible enough that people stop looking. Do not. Several boring causes produce charts that look almost identical.

Demand, cannibalisation and core updates

  • Falling demand. If branded impressions are down while position is stable, fewer people are searching for you. Branded impressions are a demand proxy. Check whether ad spend, PR or a campaign wound down around the same time, and whether the dip is seasonal.
  • Lost visibility. If position has dropped on a segment, you have a ranking problem, not an Overview problem.
  • Cannibalisation. If several of your pages share the same branded query, clicks can scatter or the wrong page can rank.
  • Core update. Impressions and average position falling together suggests an algorithm change, not a format change.
  • Other features. Changes to sitelinks, a knowledge panel or shopping units on your brand query can move clicks too.

Four ways the same chart can mislead

An impressions-up, clicks-down chart can mean several things:

  1. Query-mix dilution. New low-intent or low-position queries drag CTR down while your core terms stay put.
  2. Position illusion. Average position is impression-weighted, so it can stay flat while individual queries move in opposite directions.
  3. Title or meta mismatch. A changed title, or Google rewriting it, can cut clicks on its own.
  4. A reporting artefact. Search Console has had logging bugs before. Check Google's data anomalies page for your date range before you build a narrative.

Diverging lines are a starting point, not proof.

Get plain-English guides like this in your inbox.

One short email a month. WordPress, Shopify, SEO, no fluff. Unsubscribe in one click.

We never share your email.

If an AI Overview is confirmed on your brand

Once the data and the SERP agree, the useful question is what the Overview is doing with your name.

Audit what it says and which sources it cites

Read the Overview for each key branded query and check three things. Is it accurate? Does it link to you? Which other sources does it draw on? Outdated pricing pages, an old review site or a stale directory listing can all end up describing your business.

Citation appears to matter. Seer Interactive's April 2026 work across 53 brands found cited brands earn about twice the clicks of uncited ones. Other summaries of Seer's data give conflicting uplift figures, so we would not quote a single number. The direction is consistent, though: being the source beats being summarised by someone else. Our post on getting cited by AI Overviews and ChatGPT covers the content side.

Fix the entity signals and profiles behind it

Overviews about a brand lean on what Google believes about that brand. Tidy your organisation schema, make your about and contact pages unambiguous, and correct the third-party profiles the Overview is citing. If a directory or review site is wrong, fix it at source. We go into the mechanics in how Google understands your brand as an entity.

Reporting branded click loss without causing alarm

The worst outcome is a board seeing "branded clicks down" on a slide and concluding the brand is in trouble.

Report branded and non-branded separately, always. Use branded impressions as your demand KPI, because they show whether people still want you. Put branded conversions next to branded clicks. Our judgement, from client accounts, is that a lower branded CTR often does little commercial damage. Someone searching your name for opening hours who gets the answer from an Overview was rarely going to buy on that visit. The people who do want to transact still click through, especially when the Overview links to you.

Then state what you checked, what you ruled out, and what you are doing about accuracy and citations. That turns a worrying chart into a managed issue.

If you want a second pair of eyes on your own data, get a free website audit and we will tell you whether the drop is an Overview, a demand change or something else entirely.