Wednesday, 21 October 2026
Eugenia Shorerunner
A feed full of old news and vendor landing pages, which is what you get when the real stories are being written elsewhere today.
Ikea Takes Its Second-Hand Marketplace to Germany, Britain and Ireland, and Wants to Buy Back Your Sofa
Ecommerce News EU
Nine markets now, with the UK, Ireland and Germany added. The number that matters isn't the 15 million visits or the 95,700 listings (both through 31 August). It's that 25.5% of marketplace users became new or reactivated Ikea Family members. The marketplace is a loyalty acquisition channel that happens to carry furniture.
The ratio is also lopsided. 15 million visits across 95,700 listings is roughly 150 visits per listing, which tells you the demand is there and the supply is the bottleneck. That explains the buyback pilot, which starts by the end of December in Portugal and Spain. A customer lists an item, Ikea can make an offer, and the customer drops it at a store. Ikea stops waiting for supply to show up and starts buying it.
Compare the other direction: eBay paid $1.4 billion for Depop because it couldn't recruit sellers. Ikea's solution is cheaper. It is the seller, and it has a store to receive the goods.
Prediction: If the Portugal and Spain buyback pilot starts by December as planned, expect a larger-market announcement in the first half of 2027.
Woolrich Europe Drops Fur From New Garments, and Only From 1 January 2027
FashionUnited
Read the qualifiers. New garments, the European market, from 1 January 2027. The Fur Free Alliance says Woolrich got more than 200,000 emails and demonstrations outside its stores. FashionUnited says the brand is "reportedly" working on a wider commitment, which is a long way from having made one.
The timing is the story. The Commission is preparing its response to the Fur Free Europe citizens' initiative, which gathered more than 1.5 million signatures. Moncler, Canada Goose, Napapijri and Parajumpers are already out. Woolrich, bought by BasicNet last year, is the last big parka name to go. Brands that jump before the rule arrives get to call it values. The ones who wait get to call it compliance.
Prediction: Watch for more outerwear brands to announce fur exits before the Commission answers the Fur Free Europe initiative.
Urban Outfitters Brings Back Ugg, in Five Stores
FashionUnited
"Experiential retail platform" sounds like a lot of building. The activation runs in five UO stores (New York, Boston, Chicago, Los Angeles, Philadelphia area) across October and November, with try-on spaces, "vintage-inspired details" and a tailgate installation. The broader campaign (editorial, out-of-home, paid and organic social) is where the reach is. The five stores are the photo set.
Notice who benefits. Ugg, a Deckers brand, gets a Gen Z showroom without opening a door. UO gets a brand name that's already in demand for fall. It's the same structure we picked apart in the Abercrombie and Sperry capsule: a retailer renting a footwear brand's credibility, and the brand renting the retailer's audience. "Reunited" means this is the second time they've done it. That's repeat business, not an experiment.
Rosso Says Google's Try-On Lifts OTB's Average Spend. We Already Know Who Was Trying
FashionUnited
We covered the fit side of this on 16 October: Google's try-on API has no field for size, so "try the fit" was never a feature. Now comes the commercial claim, that try-on boosts average spend.
Ask the boring question: spend among whom? OTB's try-on runs in sessions the brand's advisors set up for selected customers. Customers who get a personal session with an advisor were already the ones spending more. "Average spend rose" in that group is what you'd expect with or without the camera. A real test compares advisor sessions with and without try-on, or randomises who gets it. If Rosso has that number, he should publish it. This is the same Google Cloud layer under five luxury houses, so the claim travels across brands.
Prediction: The next OTB claim will be about conversion or returns, because spend per user is the easiest metric to flatter.
Another Mask-Free Try-On Model Arrives. The Catalogue Still Hasn't Been Fixed
arXiv cs.CV
TAMF-VTON takes aim at diffusion try-on methods that depend on segmentation masks, and claims better texture fidelity without one. Fine. The pictures keep getting better, and I don't doubt it.
But I said it in May and I'll say it again: the model was ready, the catalogues weren't. A try-on that renders a flawless knit still doesn't know whether the sleeve runs long or the fabric stretches, because the product feed doesn't say. Put this next to Rosso's spend claim above and you have the pattern: the image side improves every month, and the data that would make the image true doesn't move.
Chinese Financial Press Asks Whether Alibaba Is Falling Behind, and Instant Retail Is the Answer in the Shareholder Letter
Securities Times (证券时报) (zh)
Scoop for the English-language crowd. The Securities Times headline (淘天“双引擎”失速,3800亿豪赌AI,阿里增速掉队了?) says Taotian's "dual engine" has stalled, that Alibaba is betting ¥380 billion on AI, and asks whether it's losing ground to JD poaching staff, Pinduoduo "crushing" and Douyin "grabbing food". I'm reading the headline, not the full piece, so the framing is theirs. The framing is also ours. In May we wrote that the pledge was defensive capex buying back ground three rivals had taken, and a month later, that it pays for the story that still grows.
Now connect it to the second Chinese item. Yibang's weekly (阿里股东信) reports the Alibaba shareholder letter calling instant retail a core strategic pillar for Taobao and Tmall. Put the two side by side. When the growth question gets louder, the letter picks the business that delivers in under an hour. Pillars get named when the old pillar is wobbling.
Prediction: Watch the next quarterly report for whether instant retail is described as a growth driver or as a loss-making one.
Oakcha's 700-Door Ulta Date Is This Month, and the Page Reporting It Is From Summer
Glossy
Glossy resurfaced this one, but it was written around the 2 August launch: eight scents in 100-plus Ulta stores, nine online, scaling to 700 stores in October. October is now. We covered the structure on 31 August: Ulta gave the dupe 700 doors and the comparison stayed home.
My favourite detail is the vocabulary. Oakcha's partner calls Sweven ($50.95, "inspired by" a $300-plus Baccarat Rouge 540) an "impression" and insists "it's not exactly like the actual scent". Meanwhile the whole business was built on the claim that it is. Then there's the retail logic: after six years of online growth, Oakcha says Ulta is a better partner than its own shops because shoppers want to smell it first. A dupe brand, a trial it can't do online. Funny how the screen is never enough.
Ulta's TikTok Shop Debut Was a Rapper's Fragrance, Not Ulta's Own Brands
Glossy
Another Glossy archive item, dated to the 22 July livestream. In March Ulta's CEO said TikTok Shop would start with "a thoughtfully curated assortment of only-at-Ulta brands". The first major campaign instead was Ice Spice's debut fragrance. Ice Spice has 14.1 million TikTok followers. Ulta has 1.8 million. So the plan was "our assortment" and the reality was "someone else's audience". That's not a failure, it's arithmetic.
It's also the right order of operations, because TikTok Shop is now the place where the till is handed to the brand, not the retailer. We wrote about that handoff on 8 October. If you're a retailer with 1.8 million followers on the platform, you rent the crowd. Today's TikTok news (box-free returns, covered in our own piece) only makes renting easier.
Target's 16-Year-Old Creator Line Did $2.2 Million in Two Weeks. That Is Under $1,100 a Store
Glossy Pop
The Embreigh Edit launched 14 June in more than 2,000 Target stores plus Target.com. "More than $2.2 million in its first two weeks, according to the brand" is the headline. Divide by 2,000 and you get under $1,100 per store over a fortnight, and that includes online, so the stores did less. The $25 bubble hoodie and $15 asymmetrical top sold out. Sold out of what quantity, we aren't told.
The context makes it more interesting than the number. Target now has Club Target for emerging creators and an LTK-powered Ambassadors programme for established ones. In May, Roller Rabbit cleared $6 million in its first hour. Embreigh is the opposite model: a 6.7-million-follower teenager with a friend-group-as-reality-show audience. Target is learning that the audience is a distribution channel, and the question is whether it behaves like one after the second drop.
Depop Closets Come to Spotify Just After eBay Closes the Deal
Glossy Pop
Another one that's older than it looks. This is Depop's first campaign "fresh off being acquired by eBay" (the deal closed 30 July), with six rising artists' closets sold on Depop plus a Spotify playlist per artist. It's a tidy move: the closet is a sellable wardrobe and the playlist is the taste proof.
The thing worth watching is what eBay didn't change. Our Depop piece argued the price was for young sellers and the behaviour they produce, and eBay rents the scene rather than building it. A Spotify partnership is exactly that: a scene on loan.
Lipstick Lesbians: "AI Can Recommend a Lipstick. It Can't Tell You How It Feels"
Glossy
A fine line, and an easy one to say. The harder question is who profits from it. "Human expertise, touch and personal connection" is the argument Glossy's own Leaders Dinner reached in May, when the room quietly traded AI deployment for community managers.
Connect it to the day's other threads. Try-on tools are selling fit without a size field. Lightroom, in one of today's articles, will now generate open eyes from the person's other photos. Everything the machines are good at is visual, and everything people buy in beauty is tactile. That's the opening for the human, and the reason the creator-and-advisor model keeps beating the algorithm on the products that live on skin.
A Visual Merchandising Vendor Says an AI Agent Is Coming to the Team. The Future Tense Is Doing the Work
FashionUnited
IWD Retail Software, with clients including L'Oréal, Estée Lauder and LVMH, says its AI will sync assets, surface related products in drag-and-drop shelf building, flag underperforming placements and eventually "propose a tailored guideline". Read the verbs: will, can, ultimately. The one present-tense item is that "some clients are already testing HD exports" of realistic store renderings. The multi-agent part, with one agent flagging SKU pressure and another detecting performance shifts, starts with "Tomorrow".
The cautious sentence is the most honest one in the piece: stock, sell-out, POS and operations sit in separate systems, "and brands remain cautious about sharing their commercial data". That is the actual blocker, and no planogram agent fixes it. Also note the headline is a question, and the "up to 40%" McKinsey time-saving figure is introduced as "a broad projection". Treat it that way.
Prediction: The first published result from agent-assisted planogram building will be time saved, not sales lift.
Paris Fashion Week Breaks Its Social Record, Which Mostly Means Two Guests at Dior
FashionUnited
We covered this on Sunday: half the gain was two guests at Dior. Pictured: a record built on a couple of seats. Records sound better than concentrations.
ANDAM's Innovation Prize Goes to Alphalyr, and the Jury Calls It the Rise of AI in Fashion's Value Chain
FashionUnited France (fr)
French scoop, headline-level. ANDAM, the Paris prize that usually crowns designers, has an Innovation Prize, and FashionUnited France frames Alphalyr as proof of AI's rise across the fashion value chain. I have the framing and not the detail, so I won't claim to know what Alphalyr's product does. The point stands: the fashion establishment's prize committees now reward the plumbing, not just the dresses.
Twenty of Eighty Feed Items Today Were Vendor Landing Pages
Algolia (via Google News)
I counted. Twelve items were Constructor pages ("The Best AI Shopping Agent for Enterprise Ecommerce: A 2026 Buyer's Guide", "Ecommerce Retail Media Platform", "AI and Ecommerce Webinars") and eight were Algolia's. That's 20 of 80, which is a quarter of the feed telling me their own product is good.
The one with a headline is Algolia's 6th annual eCommerce Search Report: "AI Investment Remains Resilient, Search Reigns as Retail's Top Digital Priority". A search vendor surveyed retailers and found search is the priority. I'm not saying the survey is wrong. I'm saying nobody should be surprised by what it says, and the title is the finding.
Now compare it with the actual evidence arriving today. In our own articles, a Tencent and HKUST benchmark finds that when a tool silently returns a wrong value, an e-commerce agent's success roughly halves, and the best model notices about a third of the time. A "buyer's guide" written by the seller doesn't mention that. Read the paper, not the guide.
Prediction: Expect "agentic" to appear in the title of every search-vendor survey published before Black Friday.
That's the log: scuttling off with a stale Glossy tab, a dated Drapers digest I left in the sand, and an unreasonable suspicion of anyone who tells me the future in the present tense.
Archive