S09.4 · Retail & Commerce
A category of categories spanning beauty, pet, apparel and electronics, where AI discovery tools erode the curation premium.
Specialty retail is a category of categories, not a single market — apparel, home, beauty, electronics, sporting goods, pet and books, each with its own unrelated market size — so no defensible aggregate figure exists. Performance splits hard by category: beauty and pet are outperforming, while apparel and electronics specialty cede share to marketplaces and mass retail. The exposed asset is the pricing premium itself. It has always rested on curation and expertise, and curation and expertise are precisely what AI product-discovery tools are now beginning to substitute for.
The largest chains sit in the US and Western Europe, and category leaders such as Sephora and IKEA operate globally. Fragmented overall, the sector is frequently winner-take-most within a category — Ulta and Sephora in beauty being the clearest example. Specialty retailers source from category-specific brand manufacturers upstream and serve consumers who want curation or expertise rather than the lowest price, which also makes this the segment most exposed to e-commerce and direct-to-consumer disintermediation. Margins range from 10-20%+ in beauty to thin in electronics, and differentiation runs on merchandising and service, not price.
Because "specialty retail" is a label rather than a single economic unit, the relevant benchmark for any operator is its own category, never the sector average. A beauty chain selling assortment breadth and in-store try-on service carries a materially different cost structure and defensibility from an electronics chain fighting marketplaces on price over identical SKUs. The categories holding pricing power — beauty, pet — are the ones where the purchase decision still depends on a service layer (application, fit, veterinary-adjacent advice) that a listing page alone does not replicate; the categories losing share sell product that is fully commoditized the moment the spec is known.
Against Amazon, curation-as-a-service and private label protect the existing revenue base, while treating the store itself as an experience is the play that can add new dollars. Both answer the same underlying risk: AI product-discovery and styling assistants erode expert curation, the value proposition that has historically justified a price premium over marketplaces.
Exposure varies by category, and the variance is the analysis. Agents optimizing directly on price and spec go straight at commodity categories such as electronics and basic apparel, where the product is fungible and the spec is all that matters — that pressure is already underway. Curation-heavy categories, where taste and fit are harder for an agent to encode, have a longer runway, with meaningful disruption expected on a 2-5 year horizon rather than immediately.