This website uses cookies

Read our Privacy policy and Terms of use for more information.

In partnership with

ChatGPT Ads turned fixed daily budgets into seven-day averages, letting OpenAI spend more when delivery looks strong. The weekly ceiling stays, but advertisers lose some control over when the money leaves.

News

1️⃣ ChatGPT puts daily budgets on a weekly leash

2️⃣ Alexa owns the shopping conversation

3️⃣ Google's auction pipe faces damages

Insights

4️⃣ AI ads have a trust problem

5️⃣ Luxury's AI front door moves upstream

6️⃣ Fashion's AI referral race

LET’S ENTER THE MAZE!

1️⃣ News

The Maze: ChatGPT Ads turned daily budgets into seven-day averages, giving OpenAI more freedom to move spend between stronger and weaker delivery days.

  • From July 27, OpenAI can spend up to twice the selected daily budget on one day, while billed media spend remains capped at seven times that amount across seven days.

  • Existing campaigns migrate automatically, and Ads Manager shows both limits; a separate campaign-total budget is stricter on total spend but does not smooth delivery across campaign dates.

  • Ecommerce teams gain more opportunity chasing around high-intent conversations, but launches and short tests can burn budget faster before conversion data proves the channel works.

Why it matters: The weekly cap protects total spend, not daily cash flow. Advertisers now set the budget while OpenAI gets more control over when it moves.

2️⃣ News

The Maze: Amazon says Alexa for Shopping’s active users nearly doubled in Q2 as interactions rose more than fivefold worldwide. The chatbot is becoming a sales gate.

  • Amazon merged Rufus with Alexa+ so U.S. shoppers can compare products, track prices, build carts and automate selected purchases across devices.

  • U.S. users spend over 40% more per order on average, while Alexa+ users join Prime at a nearly 25% higher rate, though Amazon has not proved causation.

  • Sponsored campaigns can enter the conversation automatically, giving brands high-intent exposure while Amazon controls the recommendation and measures the purchase.

Why it matters: Amazon can steer discovery, sponsored prompts and checkout in one conversation. Sellers gain conversion, but compete for fewer AI-selected slots.

3️⃣ News

The Maze: Teads is turning Google's ad-tech routing into a damages case, alleging rival exchanges collectively lost 6.88 trillion impressions over seven years.

  • Teads filed seven claims in New York, seeking damages and injunctions without naming a dollar amount.

  • Its complaint says Google Ads demand overwhelmingly enters AdX while AdX's real-time bids depend on Google's publisher ad server.

  • A Virginia court already found publisher-side monopolization and unlawful tying, but Teads must still prove its own injury and damages.

Why it matters: Auction rules mean little if one platform controls who receives the bid request; access to demand determines publisher yield and rival scale.

4️⃣ Insight

The Maze: AI advertising carries a 30-point trust penalty in France and 25 points in Britain, while the worldwide result is almost balanced.

  • Ipsos data shows every selected European market is more negative toward AI-created ad images and video than the worldwide average.

  • Globally, 79% expect companies to disclose AI use, even as 53% remain comfortable with AI targeting which ads they see.

  • EU Article 50 is narrower than `label every AI ad`; broader disclosure can still work as a voluntary trust control.

Why it matters: Brands should classify AI use, meet the legal floor, and disclose material synthetic creative when surprise could damage trust. A label cannot rescue bad work.

5️⃣ Insight

The Maze: Luxury discovery has moved upstream: 85% use AI companions, giving outside interfaces a first shot at shaping desire before brands can arrive.

  • Chatbots and AI companions reach 85% ever-used, including 52% frequent users, while AI chat or search assistants reach 78%.

  • Visual search reaches 74% ever-used, matching a wider European pattern in which AI shapes comparison and inspiration before checkout.

  • Luxury brands now need machine-readable craft, fit, and scarcity signals, plus human handoffs where identity and regret risk peak.

Why it matters: Luxury cannot win by maximizing automation. It must govern interpretation, make brand meaning legible to agents, and keep humans where confidence matters.

6️⃣ Insight

The Maze: SHEIN averaged 750,000 monthly AI referrals in H1 2026 — 1.7x H&M and nearly double UNIQLO across RETAILBOSS's selected 11-retailer comparison.

  • The top four — SHEIN, H&M, UNIQLO and Mango — captured about 86% of the 2.18 million monthly visits measured across the 11-retailer set.

  • Similarweb measured a 354.7% jump in ChatGPT homepage referrals after brand links became more prominent in May 2026.

  • Adobe found product pages averaged only 66% machine readability, but RETAILBOSS does not disclose geography, attribution or conversion.

Why it matters: AI discovery is concentrating before measurement matures; retailers must separate being named, receiving a visit and converting it into revenue.

🗞️ Quick hits

Everything else you should know

🛍️ Marketplaces rewrite access and auction rules

  • Argos launched its UK marketplace with 80 third-party sellers, expanding assortment while opening a new digital shelf for brands.

  • eBay expanded extended bidding, adding time after late bids to curb auction sniping and change seller price outcomes.

📦 Delivery becomes a conversion lever

  • Vinted added access to 18,500 DHL lockers in Germany as 70% of resellers skip in-person handover, reducing fulfilment friction for secondhand sellers.

💳 Payments and tariffs move risk upstream

  • Visa agreed to acquire behavioral-biometrics fraud platform BioCatch for a reported $2.4 billion, pulling more merchant-risk intelligence into its network.

  • Twenty-five U.S. states sued to block the latest tariffs, adding uncertainty to landed costs and cross-border pricing.

📣 Ad platforms chase cheaper intelligence

  • Meta says its new GEM training system doubled the efficiency of its large-scale ads recommendation model, lowering the cost of improving ad ranking.

  • Deliveroo cut local retail-media rate cards by 20% while launching programmatic ads in four markets, trading price for advertiser scale.

THAT’S IT FOR TODAY!

You’re the reason our team spends hundreds of hours every week researching and writing this email.

See you next time in the maze!

MarketMaze team

Reply

Avatar

or to participate

Keep Reading