That creates a very different commerce environment from a traditional ecommerce catalog, and a unique challenge for retail media and marketplace advertising.

A buyer searching for a mid-century chair, browsing contemporary art, or considering a rare watch brings different intent, context, and expectations to every interaction. Products are defined by rich attributes and complex taxonomy, inventory is highly distinctive, and discovery often extends well beyond a simple keyword search.

In that environment, relevance isn't simply a performance metric. It's part of the product experience.

As commerce becomes increasingly shaped by AI, marketplaces like 1stDibs need infrastructure capable of understanding that context and turning it into better discovery, advertising, and monetization decisions in real time. Generic ad-serving infrastructure wasn't built for that complexity.

From Ad Serving to Commerce Intelligence

As 1stDibs continued investing in its buyer and seller experience, the company reached a point familiar to many mature marketplaces: the infrastructure that helped launch its sponsored listings and retail media program was no longer the infrastructure it wanted to scale on.

The opportunity went beyond replacing an existing retail media provider. 1stDibs needed a foundation capable of supporting monetization across search, browse, and product detail pages, each with distinct discovery patterns and buyer intent.

That meant handling a large and highly dynamic catalog, supporting efficient API-based pagination, delivering fast auctions, and giving the 1stDibs team more flexibility over how relevance and monetization worked together.

The migration itself also had to preserve continuity. Roughly 1,800–2,000 sponsored listing campaigns needed to move onto the new infrastructure without disrupting the marketplace or its sellers.

Within a matter of months, 1stDibs migrated its retail media program to Topsort, including custom algorithm adjustments designed around its marketplace. On launch day, Topsort handled 100% of sponsored listings traffic.

The result was not simply a new ad server. It was a move toward a more intelligent commerce infrastructure layer.

Built Around the Marketplace, Not the Other Way Around

Luxury ecommerce makes the limits of one-size-fits-all advertising particularly visible.

A marketplace with millions of highly differentiated products cannot rely solely on rigid category structures or basic keyword matching to determine what belongs in each customer journey. Monetization has to operate within the same nuanced discovery environment as the marketplace itself.

Topsort's AI-native auction infrastructure gave 1stDibs the ability to adapt advertising to that environment rather than forcing the marketplace into a predefined model.

That includes supporting fast, real-time auctions across key surfaces while giving the teams room to continually refine relevance. Auction latency remains below 50 milliseconds, helping monetization operate without introducing friction into the browsing experience.

The infrastructure also creates opportunities to apply richer signals to surfaces such as product detail pages. Semantic matching, for example, can help identify relevant sponsored products based on the meaning and attributes of the item a shopper is viewing rather than relying only on exact taxonomy or keyword relationships.

For 1stDibs, that matters because the goal isn't simply to show more ads. It's to make monetization feel native to how customers actually discover luxury products.

Why AI Commerce Requires a New Retail Media Infrastructure

The 1stDibs story reflects a broader shift happening across commerce.

The first generation of retail media was largely about serving ads. But now commerce media is about making intelligent decisions across the commerce experience.

As catalogs become larger, customer journeys more dynamic, and AI reshapes how people discover products, marketplaces need systems that can interpret intent, products, advertiser goals, and marketplace economics together.

That's the shift from traditional ad technology toward Commerce Intelligence: infrastructure built to continuously decide what should appear, where, for whom, and at what value inside a commerce environment.

For marketplaces evaluating their own advertising stack, the lesson from 1stDibs is important. Modernizing the infrastructure doesn't have to mean rebuilding the business around a new vendor or accepting a prolonged, risky migration.

1stDibs moved thousands of existing campaigns, adapted the system to the specific needs of its marketplace, and transitioned all sponsored listings traffic within months. More importantly, it gained a foundation designed not just to reproduce what its previous system could do, but to support what comes next.

What's Next

1stDibs is already exploring that next phase: exploring new monetization surfaces, and developing more sophisticated experimentation as its program matures.

Those initiatives point toward a broader evolution of commerce media. Advertising is moving beyond individual placements and channels toward systems that can learn continuously from marketplace activity and optimize decisions across the customer journey.

For 1stDibs, the infrastructure underneath that evolution is now built to move with it. And for the industry, the takeaway is larger than a single marketplace migration.

AI Commerce requires a new intelligence layer. The marketplaces that build that layer now won't just have more sophisticated advertising programs. They'll have greater control over how discovery, monetization, and customer experience evolve together.

1stDibs is building that foundation today.