In partnership with
Happy Friday! Whatnot is tying U.S. seller commissions to four-week sales volume. Bigger merchants can earn lower rates, while smaller Premier Shops lose a discount that once rewarded service quality.
News
1️⃣ Whatnot makes seller scale pay
2️⃣ Costco keeps the membership gate
3️⃣ Walmart turns signals into action
Insights
4️⃣ AI commerce includes retailers too
5️⃣ Prime Day loses urgency
6️⃣ AI picks move brand demand
LET’S ENTER THE MAZE!
1️⃣ News
The Maze: Whatnot will link US seller commissions to four-week sales, rewarding sustained volume while removing a quality-program fee discount each month.
Existing sellers start September 21; seven thresholds set the next 28 days' category rate, from 8% down to public lows of 3.5%-6.5%.
Processing stays separate at 2.9% plus $0.30, while Whatnot charges commission on item price and eBay Live uses the full checkout amount.
The Premier discount ends November 1, so some smaller top-service sellers can pay more even as high-volume merchants save.
Why it matters: Whatnot is trading take rate for seller loyalty. The best economics now reward merchants that keep volume—and inventory—inside its marketplace.
2️⃣ News
The Maze: Costco is now on DoorDash across the U.S. Members gain another same-day channel; Costco keeps active membership as the key that opens the cart.
The launch covers all U.S. warehouses and more than 4,000 products, from groceries and bulk staples to electronics.
Shoppers select a local warehouse and link an active membership; nonmembers are routed to Costco.com to join before returning.
DoorDash gains a large weekly basket, while Costco adds reach across another aggregator without making any delivery platform its sole front door.
Why it matters: DoorDash wants restaurant users to build bigger retail habits. Costco wants more convenience without weakening the paid relationship it controls.
3️⃣ News
The Maze: Walmart is pushing Scintilla beyond reporting, using AI to explain supplier performance and route decisions into inventory, media and store work.
New AI features will surface changes, answer business questions and build dashboards, while several capabilities remain upcoming.
Existing links to Walmart Connect and Scintilla In-Store turn insights into campaigns, replenishment and field tasks.
A 2027 combined view will unite first-party and Marketplace operations, giving brands more visibility while increasing dependence on Walmart's stack.
Why it matters: Walmart can sell more than data access. It can shape the supplier decisions that follow, creating speed for brands and deeper platform control.
4️⃣ Insight
The Maze: AI could drive 15.8% of US online sales by 2030, up from 3.2% in 2026, but the forecast includes retailers' own assistants alongside external AI platforms.
The base case gains 12.6 percentage points, nearly five times the share, while intermediate percentages and dollar values remain redacted.
The definition includes retailer shopping sessions involving an embedded assistant, so AI-driven sales do not mean autonomous checkout or new demand.
Retailer-native assistants are expected to drive most of these sales, making retailer-site AI discovery relevant alongside external platforms such as ChatGPT and Gemini.
Why it matters: Separate retailer-assistant sales from outside AI referrals, then test incremental demand; a growing AI share does not automatically transfer customer ownership.
5️⃣ Insight
The Maze: Prime Day 2026 got four selling days. In IG PPC's managed-brand sample, every day still delivered a weaker lift than 2025.
Day 1 dropped from +81% to +55%, while Amazon's official event ran June 23-26 across more than 35 deal categories.
Day 4 ended at +25% versus +48% in 2025, suggesting more event length did not recreate deadline pressure.
Amazon's expanded price history makes weak discounts easier to spot, so sellers need real price depth, inventory, and external traffic.
Why it matters: Prime Day is still powerful. But it looks less automatic. Brands that treat it as free demand may just buy pricier traffic into weaker urgency.
6️⃣ Insight
The Maze: Similarweb's AI visibility study says recommended brands won the follow-on visit in six finance, travel, and beauty comparisons.
Capital One hit 14.2% versus American Express at 3.8% when AI recommended Capital One in the source visual.
Kayak reached 12.0% versus Skyscanner at 3.4%, while Sephora and Ulta each beat the named competitor when recommended.
The caveat is method: category choice, baseline switching, brand strength, location, and sample size were not visible on the report page.
Why it matters: AI visibility is not just clicks or citations. If answers shift later search and visits, brands need to measure demand created before the click.
🗞️ Quick hits
Everything else you should know
⚖️ Subscription friction gets expensive
The FTC expanded automatic Amazon Prime settlement payments and raised the maximum total refund from $51 to $200, turning subscription design failures into continuing cash costs.
📣 Retail media chases proof
Sam's Club Connect launched predictive audiences for likely future and lapsed buyers plus broader Omni-Impact measurement, selling membership-based forecasts while brands still need causal sales proof.
IAB Europe opened public consultation on in-store and travel-media measurement proposals, giving buyers clearer impression and attribution assumptions before any standard becomes final.
🛍️ Digital discovery widens the shelf
Nexi and ReFiBuy partnered to help European merchants structure product data for AI shopping discovery, moving the payments provider upstream without yet proving merchant adoption or returns.
Tesco expanded F&F's web and app range with 400 styles, more sizes, exclusives and improved search, growing the fashion shelf without matching store space.
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
