01 The consumer data
AI shoppers now close better than everyone else
The conversion gap that made AI traffic look like a novelty has flipped completely, and it took twelve months.
Adobe Digital Insights published its August AI Traffic Trends report on August 19, built on more than a trillion visits to US retail sites plus a July survey of over 5,000 US consumers. In July, retail visitors arriving from AI sources converted at a 60% higher rate than non-AI traffic.
A year ago that comparison ran the other way, and not narrowly. Twelve months earlier, non-AI visits were worth 128% more than AI visits. Travel is following the same curve without having finished it: non-AI still converts better there, but the travel gap narrowed to 1% in July.
Volume is growing fastest where the purchase is considered rather than impulsive. Adobe logged AI-driven visit share up 119% year over year in travel, 62% in retail and 31% in financial services. On the consumer side, 59% said they had used AI in the past week, up from 58% in March, and 55% say they are turning to AI more than they used to.
The finding most useful to a service business is about page type. Blogs, guides, FAQs and support pages score highest for AI citation readability across every industry Adobe measured, while dynamic content struggles. The machine reads the boring pages best.
One limit worth stating. Adobe measures retail, travel and financial services, not professional services. Read the direction of travel here, not the magnitude, and do not assume a law firm converts like a hotel.
What to do about it
Look at the pages on your site that answer a question plainly and end. A fee explainer, a what-happens-next page, a genuine FAQ. Those are the pages Adobe found machines can read, and most firms have one thin version of each buried under a services menu.
Fix the readable pages before you commission anything new. Rebuilding what you have is cheaper than a content program and it is the part the data supports.
Source
Adobe Digital Insights, “AI Traffic Trends Report,” August 2026, published August 19, 2026. Based on 1 trillion-plus US retail site visits and a July 2026 survey of 5,000-plus US consumers · business.adobe.com (PDF)
02 The research
Google trimmed the listicles, quietly and narrowly
Five million search results say the “best of” page slipped about four points at the top, and that the survivors lost the click regardless.
Kevin Indig published an analysis on August 24 covering 60,000 US-English desktop Google queries and 5.32 million organic result rows, collected in a 47-hour window this month using SE Ranking data. The question was whether Google has actually demoted the roundup pages that brands have been mass-producing to feed AI answers.
The blunt answer is no, not as a category. A listicle still appeared in the top 10 for 55.1% of queries and in the top three for 32.3%. Whether one shows up depends mostly on the question: Google returns 4.5 times more listicles when a searcher explicitly asks for a set of options.
The demotion is real but small. Comparing a fixed subset of 2,400 queries between January and August, the listicle share at position one fell from roughly 19 to 20% down to about 15%, and across the top three from roughly 36 to 39% down to 32 to 33%. Meanwhile 62% of listicles written by vendors gained estimated organic traffic since January, with the median page up 38%. Google weeded a specific kind of page rather than the format.
The number that should actually stop you is a different one. 92% of search results pages carrying a top-10 listicle also carried Reddit or YouTube somewhere on the page, and in B2B categories AI Overviews appeared on 93.4% of these queries. A page can hold its position and still lose the visit, because the answer and the discussion both sit above it.
Indig states his own limits, which is why the study is worth citing. He calls it observational, based on one crawl wave, measuring associations rather than causes, with a production holdout still outstanding as final validation. He also sells subscriptions to an audience that wants this answer.
What to do about it
If someone has pitched you a “best personal injury lawyers in your city” page with your firm at the top, ask what evidence supports the ranking beyond your own preference. Ask them separately what share of those results will carry an AI Overview above the fold.
Ranking and receiving the visit stopped being the same measurement. Any proposal priced on position alone is priced on the wrong number.
Source
Kevin Indig, Growth Memo, “Has Google demoted listicles?” August 24, 2026. 60,000 queries and 5.32 million organic rows via SE Ranking, collected over 47 hours in August 2026 · growth-memo.com
03 The local data
A specific category beats a broad one
Nearly two million business profiles suggest the general label you chose to catch more searches is costing you the top ten.
Whitespark studied 1.8 million Google Business Profiles across 4,209 categories, and Darren Shaw wrote up the findings in Search Engine Land on August 25. He discloses in the piece that he founded Whitespark, which sells local search tools.
Most owners reach for the broad category on the theory that it catches more searches. Pick Law Firm, the reasoning goes, and you are eligible for everything. The data points the other way: a specific primary category is associated with a 12.5% presence in the top 10 results against 9.2% for a generic one, roughly 36% higher.
Additional categories move the same direction and harder. Profiles carrying relevant additional categories alongside the primary ranked 6 to 17 positions better than profiles running a single primary category alone.
Shaw puts the caveat in his own copy rather than leaving it to a reader. Correlation is not causation here. The owners who bother to add categories are the same owners requesting reviews, keeping hours current and doing the rest of the work, and the study cannot separate those effects.
The reason to act on it anyway is cost. Choosing a more precise primary category takes about four minutes and carries no downside if your practice actually matches the label.
What to do about it
Open your Google Business Profile today and read the primary category out loud. If it describes a whole industry rather than what you actually do most, change it to the specific one that matches your highest-value practice area.
Then add the additional categories for services you genuinely provide, and leave off the vague catch-alls. Categories you cannot honestly claim are a listing problem, not a ranking tactic.
Source
Darren Shaw, Search Engine Land, “What 1.8 million Google Business Profiles tell us about local SEO success,” August 25, 2026. Whitespark data covering 1.8 million profiles across 4,209 categories; author discloses he is Whitespark’s founder · searchengineland.com
04 The deadline
On October 1, a missed call becomes a charge
Google is about to bill Local Services advertisers for the phone calls their office did not answer.
Google emailed Local Services advertisers this week with a policy update. “On October 01, 2026, we are updating how and when you are charged for call leads originating from Local Services Ads,” the message reads. Two changes sit inside it.
The first is the one that costs money. Missed calls during your business hours will be charged as valid leads if the caller stays on the line for more than 20 seconds. The second compounds it: if an initial call does not qualify as a charged lead, later calls between your business and that same person that meet the valid-lead criteria will be charged.
There is one carve-out worth knowing before you redesign your phone tree. If your setup asks callers to press a key to reach a department, the 20-second timer starts when they press it, and you are not charged if they never press one. Google also says it is adding safeguards against robot calls and spam abuse.
Google frames this as rewarding responsiveness, and there is a defensible version of that argument: people searching for a local professional want someone to pick up. The practical effect for a firm is different. Intake staffing just became a line on your advertising invoice.
A voicemail at 3:40 on a Friday used to cost you nothing but the client. In five weeks it costs whatever a lead in your category costs, and you will pay it again if that person calls back.
Before October 1
Pull your Local Services call log for the last 60 days and count the calls that went unanswered during business hours with the caller waiting past twenty seconds. Multiply by your average charged lead cost. That number is your new monthly exposure, and it is knowable today.
Then decide which lever you are pulling: coverage during your actual busy hours, an answering service, or a lower bid. One more thing, since this is advertising spend: your state bar advertising rules apply to Local Services Ads exactly as they apply to every other channel you buy.
Source
Search Engine Land, “Google Local Services Ads will charge for some missed calls starting Oct. 1,” August 25, 2026 · searchengineland.com · Advertiser email text reproduced by Search Engine Roundtable, Barry Schwartz, August 25, 2026
05 The money
ChatGPT starts crediting ads nobody clicked
View-through conversions arrive on OpenAI’s ad platform, and they will make its numbers look better without anything actually improving.
Search Engine Land reported on August 24 that ChatGPT Ads added three things at once: a Maximize results automated bid strategy, platform-level targeting across iOS, Android and web, and one-day view-through conversion reporting at campaign, ad group and individual ad level. The platform also expanded into Brazil and Mexico.
The view-through metric is the one to understand before your next invoice. A view-through conversion is recorded when someone converts within one day of seeing an eligible ad, with no qualifying click involved. The ad may have contributed. The person may also have been converting anyway.
The predictable consequence is that reported conversions on ChatGPT Ads will rise in September without any change in outcomes. If you compare a September number against an August number, you are comparing two different definitions.
The bidding change moves control in the other direction. Maximize results shifts bid decisions from you to OpenAI’s system, which sets and adjusts against the campaign goal you pick. Platform targeting is the counterweight, and it is genuinely useful if your conversion rate differs between app and web.
There is also a new WorkMagic integration that can push conversion signals back to OpenAI through its Conversions API. That is the same feedback loop every mature ad platform runs on, arriving at the point where the platform is also grading its own homework.
What to do about it
Split click-through and view-through conversions in every report before you compare ChatGPT Ads to any other channel. If your dashboard shows one blended number, it will show growth in September that did not happen.
If you test the automated bidding, hold something back on manual bids for a few weeks so you have a comparison that is not the platform describing itself.
Source
Search Engine Land, “ChatGPT Ads adds automated bidding, platform targeting,” Anu Adegbola, August 24, 2026 · searchengineland.com
