Japan’s competition regulator approved a plan requiring Amazon Japan to show a purchasable offer when its price test leaves the main offer slot empty. That fallback matters because it gives sellers a way to be seen when nobody meets Amazon’s price test. Once the change goes live, sellers can compare visibility with actual sales. Quiver’s acquisition of Stuart gives retailers another question to test: how many nearby deliveries can one courier serve without missing the promised windows?
News
1️⃣ Amazon Japan must show an offer when none passes its price test
2️⃣ Quiver buys Stuart’s courier network in the UK, France and Poland
3️⃣ Google makes Gemini 4 Hydrogen the default in AI Mode worldwide
Insights
4️⃣ US shoppers surveyed found shopping assistants annoying more often than helpful
5️⃣ US AI shoppers surveyed used it more for research than purchases
6️⃣ US online retail grew 12.2% in the second quarter, outpacing retail overall
LET’S ENTER THE MAZE!
1️⃣ News
The Maze: On November 11, Japan’s competition regulator approved a plan requiring Amazon Japan to show a purchasable offer when none meets its price test. The test compares each offer with the lowest price Amazon finds for the same item across a set of rival stores. Until now, missing that test across all offers could make an available item look unavailable.
Source: Japan Fair Trade Commission overview. The regulator illustrates the old empty offer area and the planned purchase option. The examples are not live Amazon screens. Browser chrome cropped.
🛒 The extra click. The Featured Offer is the seller’s offer next to the main buy buttons. When nobody cleared the price test, shoppers had to open a separate list to see what was available.
📦 One visible option. Amazon Japan must show the cheapest available offer including shipping, with a way to buy it. The display runs for five years from launch. No start date was disclosed; the plan leaves the regular price test in place.
Our take: Once the change goes live, sellers can test whether showing up without the main buy buttons is enough to earn sales without cutting their price.
📊 Visibility matters. Most purchases went through the Featured Offer, the regulator says. Measure sales alongside visibility to see what the fallback is worth.
🧾 Contracts and placement. Amazon Japan agreed to remove price parity clauses in 2017. Today’s case shows how placement can still pressure sellers to match prices, even without that old contract term.
2️⃣ News
The Maze: London delivery company Quiver bought Stuart’s parent from German investment firm Mutares. Stuart connects grocers, restaurants and retailers with independent couriers in the UK, France and Poland. Quiver wants to use that network to take its shop-to-door delivery service into more cities.
Source: Stuart via XYZ. A Stuart bicycle courier carries a branded delivery bag on a city street.
📦 The shop becomes the pickup point. The retailer can send online orders from nearby stores or small city warehouses. Quiver wants its software to group nearby deliveries so couriers can handle several orders on one route.
🏪 Two brands for now. Stuart and Quiver plan to share work on technology, sales and operations while keeping their names. The announcement gave no dates for a joint service.
Our take: The deal could give retailers another way to offer fast delivery at their own checkout. The economics still come down to what a courier can deliver in one neighbourhood.
📉 Lower prices need lower costs. Stuart largely offset price concessions by lowering its cost per order in the first half of 2026, Mutares said. Retailers testing delivery should count store packing time alongside the courier fee.
🛍️ There are alternatives. Deliveroo announced plans to deliver orders placed on Ryman’s own site or app in the UK. Compare providers on the same checkout, neighbourhoods and delivery windows. The same test works wherever a retailer has local stock.
3️⃣ News
The Maze: Google made Gemini 4 Hydrogen the default in AI Mode, the conversational version of Google Search, worldwide on November 12. Google’s Search product lead said it also handles the hardest questions in AI Overviews, the summaries that can appear above regular results.
Source: Google DeepMind. Google’s demo gives both models the same sneaker request. Gemini 3.5 Flash and Gemini 4 Hydrogen lead with different picks.
🔍 One question, several searches. If a shopper asks for breathable slip-ons with arch support, Google can look up those requirements in parallel before writing an answer.
👟 Different pages may get seen. Google says the upgrade can surface a broader range of links and show more visual comparisons across supported languages. It gave no figures on how often that happens.
Our take: Retailers should track whether their product pages show up when shoppers compare options and whether that visibility brings visitors to the store.
🗓️ Check the starting point. The core ranking update began on November 10, and a logging error reduced recorded AI impressions on November 9–11. An apparent jump from those days is a poor measure of Hydrogen’s impact.
📊 Follow the page. Search Console’s AI report shows impressions by page and country. An impression is an appearance in an AI answer. Compare those trends with store visits and orders for the same pages and markets. Google hasn’t shown how Hydrogen changes either outcome.
4️⃣ Insight
The Maze: An AI shopping assistant answers questions or helps shoppers find products on a retailer’s site. ECDB asked 3,042 US online shoppers in 2026 how they felt about these tools. The responses give retailers a reason to pay attention to where the assistant appears.
🛒 More annoyance than help. 25% rated AI assistants helpful and 37% found them annoying, rounded to whole percentages.
👥 Age groups differed. 44% found the assistant annoying in the 18–45 group, versus 30% in the 45+ group. The published labels overlap at age 45.
📊 The wider split hides variation. Subtracting the annoyed share from the helpful share, the gap was about −34 points at ages 25–34, near zero at 55–64, and +16 above age 75.
Our take: Give the assistant a specific job and test whether it helps shoppers complete it.
💬 Offer help where there’s a question. AI assistants had the lowest share of shoppers rating them annoying among the six features. Try a quiet prompt near a hard choice, such as checking whether a part fits.
🧪 Count shoppers who skip it. 38% were indifferent to AI assistants. Compare the whole visit across placements, including whether shoppers can keep browsing without opening the tool. The survey measured opinions.
5️⃣ Insight
The Maze: In a December 2025 US survey, Boston Consulting Group (BCG) found that more people who already shopped with AI reported using it for research than for purchases. A brand can make the shortlist long before the retailer sees a visit.
🔎 Finding and comparing dominate. Among US shoppers already using AI to shop, 96% use AI to find products and 80% use it to weigh options. People can report both tasks.
🛒 Fewer use AI to buy. 24% use AI for purchasing in the same group. That measures people reporting a use, rather than a share of orders.
🔐 Delegation is a separate question. BCG says 4% would let an agent buy alone. That's stated willingness, and BCG's write-up leaves the exact respondent group unclear.
Our take: Test whether an assistant can match the product and its price to the shopper's brief before investing in automated checkout.
🧾 Make the fit clear. BCG reports that about 70% use AI to compare product attributes among shoppers who use AI. Put details such as ingredients or compatibility where the assistant can find them.
🏷️ Check the price on arrival. BCG says about half use it to compare prices before buying. Run the prompt, open the listing, and check that the quoted price matches the same pack size.
6️⃣ Insight
The Maze: US online retail accelerated in the second quarter of 2026. Dollar sales grew faster online than across retail as a whole, which includes the online channel.
📈 The growth gap widened. Marketplace Pulse reports 12.2% annual online growth, versus 6.7% for total US retail. Both figures are seasonally adjusted and measure spending in dollars, not items sold.
🛒 Online took a bigger slice. Online accounted for 17.1% of retail sales, or about $17 of every $100. A larger share can coexist with growing store sales.
Our take: Compare revenue momentum with market share and order volume before moving money between channels.
🏷️ Separate prices from orders. Marketplace Pulse estimates inflation-adjusted online growth near 8%. Track orders, units and selling prices separately to see how much of your revenue gain reflects more shopping.
📏 Match the yardstick to the business. The comparison includes 28% growth in Shopify’s North American merchant sales. Benchmark a US retailer against its own online category; a platform’s sales across merchants and countries answer a different question.
🧰 Resources
Useful stuff, handpicked by the team every day
🤖 AI prompts: 10 AI Prompts for SMS & Messaging or 👉 All here
📖 Glossary: Fraud & Disputes, Decoded: 16 Terms You Actually Need to Know or 👉 All here
📅 Events: E-SHOW Madrid 2026 or 👉 All here
🔓 Get the full resource library with MarketMaze+
🗞️ Quick hits
Everything else you should know
🔍 Pinterest adds side-by-side product comparisons in the US. Shoppers can see prices, stock and reviews together before opening individual product listings.
🔀 Google Ads Editor defaults new Search campaigns to AI Max. Google can send shoppers to a different page on the advertiser’s site.
🧱 LEGO takes BrickLink marketplace offline. The site reopened November 12, and BrickLink will waive seller fees for November and December.
🔒 Shopify requires IP allowlists for custom apps on outside stores. From January, one missing list can block a partner’s next outside installation.
📦 Kickstarter sends backer orders to Shopify for shipping. Backers have 48 hours to update their address before a scheduled batch transfers.
💳 Imprint raises $130 million for retail card rewards. Kroger cardholders can earn points on eligible purchases elsewhere and redeem them at Kroger.
🗂️ Amazon Ads’ old reports face January 31 shutdown, PPC Land reports. The two reporting pages are due to become read-only December 17.
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
