
Google is dropping Gemini Tasks’ subscription barrier in six markets, while its new purchasing tools start in the US. That deserves a merchant test: permission to spend does not tell you whether an assistant can complete an order. Track completed purchases and shopper handoffs before forecasting sales. We also examine how a trusted testing tool delivered malicious code, and why bigger Temu baskets can coexist with a smaller share of orders.
News
1️⃣ Gemini Tasks becomes free in six markets; purchasing starts in US
2️⃣ Optimizely says signup flaw let attackers alter website-testing code
3️⃣ Meesho opens over 30 million products to eligible Indian YouTube creators
Insights
4️⃣ Temu loses order share across five EU markets; baskets grow
5️⃣ US TikTok shops with 1,000-plus own videos dominate estimated sales
6️⃣ Food brands get half of measured retailer clicks from paid social
LET’S ENTER THE MAZE!
1️⃣ News

The Maze: Google is making Gemini Tasks, its assistant’s action mode, free in the US, Japan, Korea, Canada, Australia and India. New purchasing and calling tools start in the US, giving merchants another possible route from a shopper’s request to an order.

Source: Google. Gemini’s Channels settings and text/email task setup; these channels start in the US.
🛠️ Assign the job. Switch Chat to Task and describe what you need. Google warns some websites still require the user’s final click.
💳 Authorize payment. Supported Mastercard and Visa cards use Google Pay; American Express comes later. Set a spending cap; Gemini is designed to obtain confirmation or advance approval before buying.
☎️ Check local limits. Calling excludes Indiana, Louisiana, Minnesota, Montana and Nebraska. Texting needs an eligible US mobile number.
Our take: Free access can widen shopping attempts. Merchants should measure completed orders before assuming those attempts become sales.
📈 A wider doorway. In July, Spark’s expansion required AI Pro. Removing that barrier in six markets changes potential reach, but tells us nothing about actual buying adoption.
🔧 Checkout remains unfinished. Universal Commerce Protocol integration into Gemini is still upcoming. Test where an order completes or hands back to the shopper; the announcement gives no completion rate or delivery date for that integration.
2️⃣ News

The Maze: Optimizely says a signup flaw let an attacker impersonate administrators and inject malicious code into Web Experimentation projects on November 16. The service tests website versions by running code in visitors’ browsers. The company says it fixed the flaw; the number of stolen cards remains unknown.

Source: Optimizely. Documentation example of the History view, with change filters and a code comparison; this is not an incident log.
🧪 Trusted delivery carried altered code. Report URI found infected snippets belonging to 178 organizations, not 178 confirmed retail breaches. Project JavaScript, custom code delivered with the tests, reached pages without a merchant changing its own website files.
💳 Payment fields made a difference. The script could read cards typed into a site’s own forms. Hosted payment frames, including Stripe and Adyen’s, isolated card fields beyond its reach. Optimizely says Feature Experimentation, its software development kits and legacy Full Stack were unaffected.
Our take: Treat testing tools as production code. Review where they run and which browser safeguards limit damage if trusted code turns malicious.
🚧 Blocking data is partial protection. Report URI found 17 organizations stopped data transmission but still allowed wallet-address substitution inside the page. Test whether harmful code can change the page, too.
✂️ Remove unnecessary exposure. Optimizely recommends removing its snippet from sensitive-input pages where no experiments run. Checkout teams should assess whether each page needs it, rather than assume every installed tool belongs everywhere.
3️⃣ News

The Maze: Meesho, an Indian online marketplace, is opening its affiliate catalogue to eligible YouTube creators. They can tag products in videos and earn commissions when viewers buy on Meesho.

Source: YouTube / Google India Blog. Official October 2024 illustration of the product shelf beneath a video; this is not a Meesho-specific interface.
🛍️ A broad shelf. The catalogue covers more than 30 million affiliate products across 40 categories from over 200,000 sellers. Tags work in videos, Shorts and livestreams.
🎥 Access has conditions. Creators need at least 500 subscribers, subject to YouTube’s eligibility rules. Meesho targets over 200,000 creators in coming months; that is a goal, not actual uptake. No uniform commission rate or sales result was disclosed. Meesho was already named as a partner in September; November adds catalogue and recruitment detail.
Our take: The useful question is which creators can bring additional buyers. Catalogue size cannot answer it.
🗣️ Local fit beats assumed reach. Meesho says 81% of its Creator Club creators come from non-metro cities and around 90% of active creators are small “nano” creators. That existing mix supports testing relevant language and audiences; it does not prove new YouTube earnings.
🏬 Attention is shared. YouTube’s India affiliate launch began with Flipkart and Myntra in 2024. Meesho joins a multi-retailer model. Sellers elsewhere can use the same test: compare completed purchases and applicable costs, without assuming this India programme is available globally.
4️⃣ Insight



The Maze: In ECDB’s five-market EU comparison, Temu, Shein and AliExpress lost order share as baskets grew. ECDB models transactions at the top 100 online stores per market. Its series spans July 2025–September 2026; the EU parcel duty began on July 1, 2026.
📉 Order share fell. ECDB reports that the three marketplaces lost 11 percentage points of transaction share in ECDB’s rounded before/after comparison, whose exact windows are not specified. The UK share stayed near 10% after July.
🛒 Temu’s baskets grew. Temu’s average order value rose 37% in July, meaning spend per completed order. The UK and US saw no similar July jump. The timing does not prove the duty caused the change.
🏷️ Rivals saw smaller gains. Shein’s average order value rose 14% and AliExpress’s rose 1%. ECDB does not specify the exact baseline for Temu’s 37% July increase or the periods behind Shein’s 14% and AliExpress’s 1% gains. Their order shares fell from 5% to 3% and 2% to 1%, respectively.
Our take: A bigger average basket can conceal the loss of cheap orders.
📊 Measure the whole channel. Temu’s order share fell from 16% to 8%. Share alone cannot establish an absolute order decline. Track order count and margin alongside basket value.
📦 Test bundles by category. The duty is €3 per product category in qualifying parcels under €150. A bundle within one customs category can spread one charge across more units. Mixing categories can trigger extra charges; test what shoppers will pay.
5️⃣ Insight

The Maze: US TikTok Shop sales are concentrated among shops publishing large volumes of their own videos. Marketplace Pulse, a marketplace research firm tracking more than 150,000 US shops, compares each video-count band’s share of sellers with its share of estimated sales.
📊 A small group dominates. Shops in the 1,000-plus-video band account for 8.4% of sellers and 60.3% of estimated sales. That is concentration, not proof that posting more causes sales.
🎬 Low output takes little share. Accounts with 50 videos or fewer make up 38.6% of sellers but 4.7% of sales. These totals add the three lowest bands.
Our take: Brands evaluating this channel should budget for ongoing production, then test whether the work pays. A posting quota alone is a weak business plan.
⚖️ Test the economics. The 201–1,000-video band contains 29.4% of sellers and 29.3% of sales. Near parity offers no universal upload recipe; compare production cost with profit.
🤝 Count both kinds of promotion. Own-account counts exclude affiliate creator content and are available for 71% of tracked sellers. Assess brand and outside-creator activity separately. The sales window is unspecified, and established shops may have more resources to produce videos.
6️⃣ Insight

The Maze: Half of the observed food and beverage retailer click-throughs came from paid social by Q3 2026. That made one route to retailers more important inside MikMak’s measurement network.
📱 Social took a larger share. MikMak reports paid social’s rise from 36.7% in Q1 to 51.0% in Q3. It counts sessions in its measured food and beverage content with at least one click to a retailer, rather than completed purchases.
🔍 Search lost share. MikMak’s channel figures imply combined search falling from 48.8% to 35.9% after adding paid and organic search. They do not show whether the number of visits from search fell.
Our take: Look beyond the click share before committing more of the media budget to social.
🗓️ Check when the change happened. With paid social already at 51.3% in Q2, the big move came before the final quarter shown. Check what changed in campaign budgets, products and retailer destinations before calling it a sustained acceleration.
🛍️ Audit the retailer landing. Even with paid display rising from 4.2% to 6.9% between Q2 and Q3, social remained the largest measured route. Test the two routes on completed orders and media cost, including cases where a retailer is out of stock.
🧰 Resources
Useful stuff, handpicked by the team every day
🤖 AI prompts: 16 AI Prompts for Forecasting, Cash Flow & Budgets or 👉 All here
📖 Glossary: APIs & Integrations, Decoded: 18 Terms You Actually Need to Know or 👉 All here
📅 Events: ECBR Summit Plataformas 2026 or 👉 All here
🔓 Get the full resource library with MarketMaze+
🗞️ Quick hits
Everything else you should know
🏦 Temu adds Pay by Bank in Germany. TrueLayer also provides seller bank-account verification across eight European markets; wider European checkout availability is planned.
🛒 Airbnb adds Deliveroo groceries in five cities Guests in London, Manchester, Edinburgh, Milan and Dubai can order through Airbnb up to five days before check-in.
👗 Daydream’s AI stylist can now complete purchases The company says Daydream Agent remembers shoppers’ style, fit and budget and offers one-click checkout.
💸 Google raises France and Italy ad surcharges to 3% From January 1, 2027, fees follow where ads run, even for businesses based elsewhere.
🛒 Allegro’s Polish marketplace sales rise 14.7%. The Q3 year-on-year increase measures goods sold through the platform; its Polish revenue rose 17.1%.
⚖️ EU flags gaps in Temu’s product checks The findings remain preliminary. The Commission says Temu often fails to verify required product-test reports.
👟 Nike targets about $5bn extra revenue by fiscal 2031. The plan uses fiscal 2027 as its baseline and calls for assortments tailored with retail partners.
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

