Article
Why Local Businesses Should Track AI Overview Searches Separately
A Pew Research study reported by Ars Technica found Google AI Overviews cut website referrals nearly in half, which is a reason to stop reading one blended traffic number and start splitting it by query type.
Owners of independent local service businesses in the United States
- Supporting sections
- 4
- Questions answered
- 3
- Research captured
- 2026-09-24
If you run a plumbing company, a salon, a law office, or any other local service business, you probably check your website traffic the same way you always have: one total number, maybe broken out by month. That worked when search results were a plain list of links. It works less well now that Google shows AI-generated summaries, known as AI Overviews, directly on the results page for many searches.
A Pew Research study of 900 US adults, reported by Ars Technica in August 2026, found that Google AI Overviews reduced website referrals by almost half compared with the traditional "10 blue links" format. That is a significant number, but it describes what happens on searches where an AI Overview appears, not your traffic as a whole. If your dashboard only shows one blended total, you cannot tell which of your target searches are affected or by how much.
This piece explains what the finding says, why it matters for a business that depends on being found in local search, and a specific test to run on your own numbers instead of guessing.
The argument
The thesis
Because AI Overviews appear to reduce referral clicks specifically on the searches where they show up, local business owners should stop relying on a single blended traffic total and instead track impressions, clicks, click-through rate, and referral activity separately for queries that display an AI Overview versus those that don't.
The case
What the research shows
What the study actually found
The finding comes from a Pew Research study of 900 US adults, as reported by Ars Technica in an August 2026 article about Reddit's CEO questioning the value of Google's AI Overviews. According to that reporting, Google AI Overviews reduced website referrals by almost half compared with the traditional format of ten ranked links. The article does not provide a category-by-category or industry-by-industry breakdown, so it should be read as a general finding about search behavior rather than a number specific to local service businesses.
Why a blended traffic number can mislead you
Most small business owners look at total organic sessions or total leads for the month. If AI Overviews are reducing referrals on some searches but not others, a blended number can stay flat or even grow while the searches that used to reliably send you clicks quietly decline underneath it. You would have no way to see that shift, and no way to know whether a marketing change you made helped, hurt, or was simply masked by a shift in how many of your target searches now trigger an AI Overview. Splitting your monitoring by query type is the only way to see what's actually happening to the searches that matter most to your business.
The proposed test: baseline, annotate, compare
The recommended approach has three steps. First, for the specific queries that matter to your business, identify which ones currently display an AI Overview and establish a baseline for impressions, clicks, click-through rate, and referral activity on those queries. Second, annotate when you observe an AI Overview appearing on a given query, since this can change over time as Google adjusts which searches trigger the feature. Third, compare performance over time on those annotated queries, without assuming that any change you see was necessarily caused by the AI Overview itself. This is a monitoring discipline, not a guarantee of a specific outcome: it gives you the data to see what changed, not a conclusion about why.
Takeaway
A single blended traffic total can hide the effect AI Overviews are having on your most important searches. The fix is not a new tool or a big budget item, it's a monitoring habit: separate the queries that show an AI Overview from the ones that don't, and watch them individually.
Step 1
Baseline your AI Overview queries
Identify which of your target search terms currently show an AI Overview and record current impressions, clicks, click-through rate, and referral activity for those terms specifically.
Step 2
Annotate changes as they happen
Note when a query starts or stops displaying an AI Overview, since Google's use of the feature shifts over time and your baseline needs to reflect that.
Step 3
Compare without assuming causation
Review the annotated queries over time and treat any change as a data point to investigate, not proof that the AI Overview caused it.
Questions
Frequently asked
Keep reading
Related
1 sources, captured 2026-09-24 · show· hide
Vendor pages change. Figures are as published on the capture date; confirm before deciding.