The Watson Weekly Weekend edition is sponsored by Avalara - the agentic AI platform automating global tax and compliance for leading eCommerce brands.

Unilever cut 2,200 agencies and saved $500M. P&G saved $750M in-housing media. GLP-1 now leads to brides receiving legal waivers with their dresses and a European fast fashion brand picks a US marketplace.

It's the first time the Swedish fast-fashion giant has listed a product on any US online marketplace, and out of every option, they chose Nordstrom. Not Amazon. Not Walmart. Not Target. A department store with a fashion-forward customer and a marketplace that is quietly working better than almost any new entrant out there. Price points run $20 to $50, with one $400 leather jacket I assume was made in a run of twenty.

This is a brand credibility play. H&M wants to be seen next to Vince and Madewell, not next to a third-party seller out of New Jersey. Nordstrom gets selection and volume with zero inventory commitment. Every apparel brand I've talked to about Nordstrom's marketplace says the same thing. Strong demand. High-quality customers. The pattern is clear: marketplaces bolted onto real retail brands with real distribution work. Standalone marketplaces mostly don't.

New data says 10% of couples planning weddings this year are on a GLP-1, and over half started specifically for the wedding. Some users lose a clothing size every two to three weeks. A wedding dress takes six to nine months and costs thousands.

Bridal shops have started requiring waivers. A product designer in NYC lost 50 pounds, picked a gown three sizes smaller than her current size, and signed paperwork that said the shop wasn't liable if the dress didn't fit on the day. I don't blame the shops. They're protecting 80 hours of tailoring against a variable they can't control. The bigger story is what GLP-1 does to apparel demography over the next five years. Plus-size assortments built around a 67% market are going to look very different when that number moves. Nobody knows how much it moves. That's the part nobody's modeling yet.

Unilever cut its agency roster from 3K to under 800 and saved $500M. P&G in-housed media for $750M in fee reductions. A retail season used to be a quarter. Now it's two weeks. If you're still planning by quarter, you've already lost a turn.

The merchant role survives but contracts. The 2026 merchant isn't editing product. They're editing algorithms, owning a category, and running AI tools that used to take a team of four. Fewer merchants. Better ones, if they can hold a point of view and read a dashboard. Founder-led brands have operated this way for years. Everyone else is catching up.

The thread I haven't worked out: if we skip the associate buyer and assistant buyer rungs, who teaches the next generation taste? AI runs the numbers. Taste is still human, and taste is taught on the job.

Open question. I'll keep watching.