Temu’s EU plan faces its next test

Test the safeguards. Temu must assess risks specific to its EU marketplace and use reliable measures to check whether its controls reduce the spread of illegal content.

Examine how offers spread. Temu must review moderation weaknesses, software that recommends products, and ways traders might exploit marketplace features to spread illegal listings.

Implementation precedes proof. Temu has four months to implement the plan, followed by an independent audit. Temu welcomed the approval; no completed audit or measured safety improvement is established.

A lesson beyond Europe. Removing an illegal listing and understanding how recommendations spread it are different jobs for marketplace operators. This decision concerns Temu’s EU service.

Keep the stages separate. A plan, its implementation and an audit answer different questions. Treating approval as the outcome would skip the evidence the process still needs.

One parcel can trigger several fees. Declaration lines list separate goods categories. Dutch Customs' example puts two shirts, three sweaters and a hat on three lines: €6 in handling fees.

Duty still comes on top. The existing €3 duty applies per distinct item on eligible shipments up to €150. Above €150, normal duties replace that flat rate; the handling fee remains. Private gifts and ordinary business-to-business trade are excluded.

Checkout terms decide the bill. The fee funds customs checks. Businesses pay customs and may pass the charge to shoppers at checkout or delivery. Estonia says declarations submitted from 1 November qualify even for earlier orders.

Stock location matters. Irish Revenue says goods already in the EU when ordered avoid this direct-import charge. Earlier import and storage costs remain.

Demand effects remain unknown. The fee starts in November. No measured effect yet proves that shoppers will switch or sellers will relocate inventory.

Convenience needs a lasting supplier. Caper’s cameras, sensors and scales identify groceries while shoppers fill their carts. A screen tracks their basket, reducing manual scanning. The wind-down ends the business supplying that equipment.

Two milestones differ. Instacart reportedly aims to remove carts by March 2027. Its filing targets substantial overall closure by June. Individual store schedules and contract remedies remain unclear.

The bill includes noncash costs. Forecast pretax closure charges are US$55 million–US$80 million, covering contracts, asset-value reductions and employee termination. Expected future cash spending is US$30 million–US$50 million. Instacart retains other in-store software.

Test both sides. A store pilot tests customer behaviour. It does not guarantee the vendor will keep funding the product.

Keep the conclusion bounded. Instacart has not disclosed Caper’s full economics. Its exit cannot establish that all smart carts fail.

The day-1 leader held the budget crown 45% of the time, then jumped to 60% on day 2 and 67% on day 3.

The blind-guess baseline was about 20%, so the day-3 read was already more than three times random selection inside the launch batch.

The caveat is scope: mostly lowest-cost bidding across mixed CBO and ABO launches, so vertical, campaign type, and persona breadth still matter.

Financial filings beat CEO statements by 12.4 percentage points in the survey.

Current employees reached 15.2% and former employees 13.3%, making staff voices stronger credibility checks than the corner office.

An earlier study found 78% say a CEO affects their view of the company, so visibility still matters—just not without proof.

Cisco was the world's most valuable company in March 2000, yet the FT/LSEG data marks recovery around 25 years later.

Amazon reshaped retail and cloud, but the visible recovery note still says it took 10 years after its December 1999 peak.

Qualcomm needed 20 years, while Nvidia recovered by October 2006, proving the rule: winner quality and entry price are separate jobs.

🤖 AI prompts: 14 AI Prompts for Google Search Ads or 👉 All here

🛠 Tools: 12 Best Ecommerce Shipping Labels & Delivery Management Software (2026) or 👉 All here

📖 Glossary: Omnichannel & Store Operations, Decoded: 15 Terms You Actually Need to Know or 👉 All here

📅 Events: D2C Summit 2026 or 👉 All here

Venmo grew payment volume in the mid-teens as PayPal checkout growth stayed near 1–2%. US debit-card use can rise without more shoppers choosing PayPal at checkout; the two need separate evidence.

Visa confirms commercial digital-dollar funding for eligible cross-border payout firms through Inyo. Recipients can still get local currency; Visa Direct’s 23% growth measures the whole service, not stablecoin use.

UPS handled fewer U.S. parcels but earned more per package; fuel charges supplied about half the 6.1% revenue-per-parcel gain, limiting how much the headline alone says about a stronger customer mix.

OpenAI launched GPT-6.1 Terra with standard developer usage prices 50% below GPT-5.6 Terra and GPT-6.1 Sol. Cheaper processing is one cost component; savings still depend on errors, retries and review.

A vendor found two AI mentions in copy for 3,379 flagged Facebook and Instagram ad files. Its search missed on-screen notices, so this is not a count of every disclosure.

US regulators and twenty states sued Dollar General over alleged checkout overcharges. The disputed case tests whether printed price labels actually reach shelves; a court has yet to decide liability.

EU authorities detained a record 196 million suspected counterfeit items in 2025. But 16 million tobacco items held at the border were later released; detention counts are not completed seizures.