What a billion-dollar commerce desk actually costs

Wirecutter is the reference implementation of publisher commerce, and the numbers are public enough to learn from. The New York Times bought it in 2016 for a reported $30m-plus. By 2018 it was returning more than $20m a year. In 2023 and again in 2024 it drove over $1bn in gross merchandise value through its recommendations.

That last number is the one that gets quoted in board decks, and it is the wrong one to plan against. GMV is what readers spent at retailers, not what the publisher kept. The commission on it is a single-digit percentage, and affiliate revenue at the Times is reported inside a bucket called "other revenue" alongside syndication and licensing, so the exact take is not separately disclosed.

The cost line nobody quotes

Wirecutter runs a newsroom of 160 journalists. That is the actual barrier to entry, and it is why most publishers who announce a commerce strategy quietly stall a year later. The content is not the hard part in isolation; the hard part is sustaining testing-grade output across enough categories, for long enough, to compound.

It is also why Wirecutter has spent the last two years diversifying away from pure affiliate. After operating solely on commissions for years, it added advertising, and hit its 2025 advertising target by midyear with hundreds of advertisers running campaigns on the site.

What it means if you aren't the Times

You are unlikely to fund 160 journalists to find out whether category 400 was worth entering. The alternative is not lower ambition. It is a different cost structure: buy the category intelligence and the operational spine, and keep your editorial judgment where it belongs.

That is the entire premise of a desk model. Wirecutter proved the ceiling is high. It also proved the fixed cost of reaching it the traditional way.