SEO Recap covering July 24, 2026
Daily summary of what matters in SEO, GEO, AEO, and AI search generated with Claude Code (beware of hallucinations)
Generative Engine Optimization (GEO) & Answer Engine Optimization (AEO)
The 84% "identity leak": why AI search can't verify your business
Source: Search Engine Land
- An audit of 71 verified businesses on Prince Edward Island found the average business resolved only 15.6% of what an AI system needs to trust and retrieve it, leaving about 84% of its identity unverifiable. 17% had no AI-retrievable presence at all.
- The author names the most common failure modes:
- Real leadership and history buried on secondary pages an AI never reaches (22 of 71 had named leadership on their own site)
- Modern sites built entirely in client-side JavaScript that return zero extractable text to a direct fetch
- Dead or lapsed domains, and identities split across multiple domains or directory listings with no canonical source
- The audit used a condensed E-E-A-T scoring framework (485 points) plus checks for NAP consistency, an identifiable named human, and whether a direct fetch returned readable text.
- The recommended fixes are low cost:
- Put a real named person on a page and link it from the homepage
- Server-render core facts (name, address, services, leadership) even if the interactive site stays client-side
- Consolidate to one canonical domain and confirm it is live
- Add a linked privacy policy and terms page
- Check whether a third-party reseller outranks your own booking path
Emergency brand audit: what AI says about your locations
Source: Search Engine Journal
- GatherUp's Annie Jackson and Jason Wertham showed a query for a "no-touch car wash for my SUV in Norfolk, VA" returning a 3.3-star business with clearance and hours answered above the rating, showing query match can outrank star rating in AI answers.
- GatherUp consumer data from fall 2025: 55% had consulted Google or Bing AI summaries, 48% had asked ChatGPT about a local business, and 31% had asked multiple times.
- Google, Yelp, and other major directories block LLMs from crawling review content on listings, so the presenters advise republishing reviews to social channels or on-site widgets to make them crawlable for AI answers.
- Review recency and velocity beat average star rating: 45% of users prioritize recency over rating, 60% trust detailed written reviews over rating-only, and 70% prefer a review request within 72 hours.
- The presenters, citing SparkToro research that identical LLM queries never return in the same order, recommend running audit prompts in incognito or temporary-chat mode and re-running monthly, treating total citations rather than position as the visibility metric.
The AI Overviews YouTube gap most SEO teams skipped for 20 years
Source: Search Engine Journal
- YouTube's 2025 U.S. Impact Report, built on Oxford Economics research, says the platform's creative ecosystem contributed more than 60 billion dollars to U.S. GDP and supported over 540,000 full-time equivalent jobs, with 76% of small and medium businesses saying it helped grow their customer base.
- The author argues YouTube videos increasingly surface inside AI Overviews as the primary cited answer, making a long-deprioritized channel a reliable path to AI citations.
- Rather than building an in-house channel from zero, the author recommends a partnership strategy:
- Identify creators with genuine authority in your specific category, not the biggest names in budget range
- Fold influencer partnerships into SEO, content, and social workflows instead of a separate line item
- Measure the downstream referral and citation effect, not just video view counts
The real risk in agentic commerce: legibility isn't preference
Source: MarTech
- The piece distinguishes algorithmic legibility (structuring product data so an agent can read and rank a brand) from brand preference (a human asking for a brand by name), warning that a brand can be readable to the machine yet absent to the person.
- OpenAI's Instant Checkout went live in September 2025 and was pulled in March 2026 with only about a dozen merchants integrated, while Google's Universal Commerce Protocol, announced at NRF in January 2026 with Shopify, Etsy, Wayfair, Target, Walmart, and Visa, continues to expand.
- Salesforce found 39% of consumers, including 54% of Gen Z, have used generative AI to discover and evaluate products, and 63% of Gen Z are interested in AI agents making purchases, though that reflects stated interest rather than measured behavior.
- SparkToro found less than a 1% chance the same brand appears across two identical AI queries, which the author uses to argue brands need zero- and first-party preference data they control rather than static product feeds.
- The author recommends action:
- Audit the zero- and first-party data you actually capture, especially why customers chose you
- Build relationship data in owned channels: discovery, community, the delivery moment, and a direct channel
- Put a real number on the "discovery tax" and place ownership of agent-facing data inside marketing
Technical SEO
Google bans fake or undisclosed incentivized reviews in review snippet structured data
Source: Search Engine Land
- Google added a new guideline to its review snippet structured data documentation telling sites not to include fake or undisclosed incentivized reviews on the page or in markup.
- Google spelled out what counts as fake or undisclosed:
- Reviews not based on a genuine experience of a product or service
- Reviews written in exchange for money, discounts, vouchers, or free products without a clear and prominent disclosure
- To stay eligible for review rich results, Google says publishers should remove any reviews obtained through these methods.
7 mistakes that undermine your SEO tests (and how to fix them)
Source: Search Engine Land
- The author, citing a decade of testing with a 70% average success rate, calls incrementality testing (test pages with a change versus a control group over the same period) the gold standard for isolating ranking impact, versus A/B testing (better for UX and CRO) and pre/post testing (fast but unreliable for SEO).
- A rigorous SEO hypothesis should be:
- Actionable: a change big enough to matter with enough sessions to confirm
- Consistent: applied across multiple pages to rule out a fluke
- Measurable: tracking and performance data in place
- Extensive: run long enough to see ranking and visibility impact
- The author recommends risk mitigation before going live: QA across browsers and devices, launching on one page first, no Friday launches, avoiding highest-value pages, weekly checks, and a revert plan.
- When reading results, the author warns to check all data, validate surprising numbers against another source, look past surface traffic to keyword and lead quality, segment by device and user, and watch for outliers skewing the outcome.
Organic Search & Algorithm Updates
Why a "wait and see" SEO strategy carries its own risk
Source: Search Engine Land
- The author argues that pausing SEO amid AI search uncertainty creates a momentum deficit, since organic trust and authority compound over time and cost more to rebuild than to maintain.
- The piece recommends staying focused on fundamentals that support both search and AI visibility: original content, data, E-E-A-T, and full-funnel content, rather than chasing every new tactic.
- The author suggests using "trigger events" (shifts in internal priorities, competitor moves, or major algorithm and LLM changes) to prompt strategy reviews instead of a full stop.
- The author cites Lily Ray's data suggesting popular listicle strategies may be backfiring, with some authors left off their own industry lists, as an example of why continued testing matters.
SEO Pulse: search growth slows as publishers weigh leaving Google
Source: Search Engine Journal
- Alphabet reported Q2 2026 Google Search and other revenue up 17% year over year to 63.27 billion dollars, easing from 19% in Q1, the first slowdown after four quarters of acceleration. Pichai pointed to Gemini 4, now in pretraining, and Alphabet raised 2026 capital spending guidance to 195 to 205 billion dollars.
- A federal judge dismissed Google's DMCA anti-circumvention claims against SerpApi, ruling that bypassing an access control on public search results is not circumvention when the results contain no copyrighted content. Google may amend narrower claims involving licensed images within 21 days.
- Major publishers are weighing their presence in Google's index: USA Today Co. says it may delist within 6 to 12 months, Reddit is reevaluating its 60 million dollar per year licensing deal, and Politico, Reuters, and People Inc. are considering crawler limits or blocks.
- The EU issued its first DMA fines against Google on July 23, totaling 890 million euros (460 million for self-preferencing, 430 million for restricting developer steering), a week after decisions requiring Google to share anonymized search data with rivals and open Android to competing assistants.
Bing tests fading out related searches on hover
Source: Search Engine Roundtable
- Microsoft is testing an effect in Bing search results where hovering over one related search highlights it and blurs the other suggestions.
- The behavior was spotted and shared on X by Khushal Bherwani and Radu Oncescu.
Published on