In partnership with
China is separating credit from checkout: from September 30, payment apps cannot present loans as payment options or bundle financial products behind default consent. The checkout may still close the sale, but lenders will have to win a more deliberate borrowing decision.
News
1️⃣ China separates credit from checkout
2️⃣ McDonald’s puts waiting time on sale
3️⃣ Banks want safer AI shopping
Insights
4️⃣ Shein sells more, keeps less
5️⃣ AI value needs connected levers
6️⃣ Vinted hired beyond marketplaces
LET’S ENTER THE MAZE!
1️⃣ News
The Maze: China’s September 30 rules separate credit from payment options and block default consent for bundled financial products, changing how apps acquire borrowers.
Non-bank payment institutions cannot list loans as payment options, while installment plans can remain available under the wider rules.
Authorized platforms must route customers to the financial provider’s own platform, with clear identity and a mandatory reading pause before purchase.
Removing bundled-product defaults makes borrowing a separate decision, potentially changing lenders’ customer-acquisition economics without proving a fall in sales.
Why it matters: Checkout can still close the sale, but borrowing needs a clearer choice; platforms and lenders must watch whether deliberate credit uptake replaces default-driven demand.
2️⃣ News
The Maze: McDonald’s is testing other brands’ ads on US restaurant screens after customers order, turning the wait for food into a potential new source of advertising revenue.
The limited pilot puts ads on drive-thru boards and self-service kiosks after ordering at selected company-owned US restaurants.
More than 13,700 US restaurants offer potential scale, but the confirmed test has no disclosed nationwide rollout or ad revenue.
US comparable sales rose 0.8% in the second quarter, making income from existing visits worth testing without asking diners to buy more food.
Why it matters: McDonald’s could earn twice from one visit: a meal sale and an ad. The test must prove that advertiser value outweighs extra costs and customer irritation.
3️⃣ News
The Maze: Six banks, including Bank of America and ING, have proposed safeguards for AI-assisted shopping as payment technology moves ahead of customer protections.
The banks published principles covering transparency, safety, privacy, choice and how different services work together.
AI purchases raise questions about spending authority, card-data security and who handles mistakes when a customer delegates the buying.
The framework is voluntary, with implementation still to come, so merchants cannot yet treat it as a finished payment standard.
Why it matters: Merchants need proof that an AI order reflects the shopper's instructions and a clear route for disputes before faster checkout becomes dependable revenue.
4️⃣ Insight
The Maze: Shein's 2025 sales reached $41.85 billion, but earnings fell to $2.06 billion, leaving the larger retailer with much less profit from each revenue dollar.
Revenue rose 8.0% while net income fell 38.7%, taking the annual net margin from about 8.7% to 4.9%.
The 2023–2025 figures show that sales scale expanded faster than retained profit, reducing the earnings cushion available to absorb higher costs.
Q1 2026's $99 million loss included a $328 million accounting charge, so its reported result needs that qualification and a separate quarterly comparison.
Why it matters: Revenue growth can leave a thinner earnings cushion; operators need to explain how sales become profit and what caused the margin to deteriorate.
5️⃣ Insight
The Maze: AI creates material value when growth, productivity, supply-chain execution, and margin improve as one operating loop rather than unrelated pilots.
Productivity carries the largest stated range at 30–50% in core commercial and support activities.
Growth at 10–15% requires relevant offers that also reflect price, availability, and delivery reality.
The four ranges are separate benchmarks, not an additive ROI claim.
Why it matters: The better roadmap connects customer signals to trading, inventory, fulfilment, and service. Each improved decision should make the next one cheaper and smarter.
6️⃣ Insight
The Maze: Only 38.8% of Vinted's mapped product leaders came from marketplaces; the company protected domain expertise without making it a hiring monopoly.
In the 151-person sample, 38.8% had marketplace experience, 14.5% ecommerce experience and 46.7% neither.
Marketplace backgrounds rose from 19% before 2016 to 41% after, suggesting Vinted became more deliberate about two-sided platform pattern recognition.
Banking, ad-tech, telecom and Booking.com talent added payments, data and scale capabilities around the marketplace core.
Why it matters: Hire marketplace fluency where network effects bite, then widen the aperture for payments, trust and logistics; category pedigree alone is too narrow.
🗞️ Quick hits
Everything else you should know
🤖 Agentic payments move from principles to proof
Danske Bank and Mastercard completed Denmark’s first payment initiated and finished by an AI agent inside a regulated bank, turning agentic commerce from a standards debate into a live authorization and dispute test.
⚖️ Platforms absorb more checkout and ad risk
A Frankfurt court treated Meta’s ad-auction role as active control rather than passive hosting in a fake-ad case, exposing the platform to penalties reported at up to €250,000 per breach.
Shopify expanded Shop Pay Installments across multiple business entities in the US, Canada and UK, removing the primary-entity limit for eligible multi-market merchants.
🛒 Retailers pull demand and checkout forward
Amazon, Walmart and Target scheduled closely overlapping October sales events in the US, forcing sellers to fund inventory, discounts and advertising across three channels at the start of the holiday cycle.
Lidl is rolling Scan & Go into France through its Lidl Plus app and plans it as standard in every new French store, making app adoption part of checkout infrastructure.
📻 Retail media gets transaction-level proof
Experian and Bauer Media reported more than 20x return on ad spend for an Asda audio campaign, plus 13% higher spend and 27.5% higher retention versus a control group, giving buyers a strong result that still needs methodological scrutiny.
THAT’S IT FOR TODAY!
You’re the reason our team spends hundreds of hours every week researching and writing this email.
What do you think of this issue?
See you next time in the maze!
MarketMaze team
