Yahoo's Brickhouse: a startup factory inside the corporation, gone in two years

By 2007 Yahoo was losing ground to Google and its best people were leaving to found startups. Brickhouse was the answer: an internal incubator away from the Sunnyvale campus where employees could build new products at startup speed, with none of the startup risk.

Recreate the startup inside the corporation: a separate SOMA space, four-to-six-person teams on four-to-six-month sprints, corporate pay and benefits, cheap failure.

Yahoo built a startup-styled space in San Francisco's SOMA district, run by Salim Ismail under Bradley Horowitz, with Flickr co-founder Caterina Fake as a visionary. Teams of four to six took an idea for four to six months and shipped it under the Yahoo brand. In two years Brickhouse launched Pipes, a visual web-mashup builder; Fire Eagle, a location platform; Yahoo Live, a livestreaming service; and the game BravoNation.

The products outlived the unit. Pipes defined the mashup era, Yahoo Live anticipated Periscope and Facebook Live by years, and Fire Eagle got to location services early. But the unit burned through leaders — Ismail and Fake left early on, and successor Chad Dickerson departed for Etsy in mid-2008. In December 2008, with Yahoo in crisis, the company announced Brickhouse would close by year's end.

In early 2007 Yahoo opened Brickhouse, an internal incubator in San Francisco's SOMA district, far from its Sunnyvale campus. The pitch: recreate the startup environment — small teams, short projects, a space that felt nothing like a big company — with the full backing of a giant. As one early visitor put it, roughly four to six people would work an idea for four to six months, on salary and benefits, and a failed idea was no big deal: the team simply rolled onto the next one or returned to headquarters.

The design was also a retention play. Talented employees who would otherwise quit to found a company could build their idea inside Yahoo instead. And Brickhouse shipped: Pipes, an ambitious visual way to program the web; Fire Eagle, a location-brokerage platform; Yahoo Live, a live video service years before Periscope or Facebook Live; and the social game BravoNation. For a moment Yahoo was, in Variety's words, the place where the cool kids wanted to work.

It did not survive its parent. Leadership turned over constantly — Salim Ismail and Caterina Fake were gone early, Chad Dickerson left for Etsy in 2008 — and as Yahoo's own crisis deepened, the company announced in December 2008 that Brickhouse would shut by year's end. TechCrunch's verdict: innovation should be embedded in the fabric of a company, not outsourced to a small group of people in San Francisco.

Cheap experiments: with salaries and benefits covered, a failed four-to-six-month project cost little — teams rolled onto the next idea or back to Sunnyvale.

Distance mattered: a separate space away from campus let people work like a startup, without feeling like part of a big machine, while keeping the parent's distribution.

It answered attrition: would-be founders could chase their idea inside Yahoo with corporate pay, instead of quitting to chase it outside.

The products proved the model — Pipes, Fire Eagle and Yahoo Live were years ahead of the market; the unit innovated even as the company around it did not.

An innovation unit can ship faster than its parent, but it cannot fix the parent — and if the parent hits crisis, the unit is the first thing cut. Isolate innovation and you isolate its effects.

Variety's 2016 retrospective made Brickhouse the emblem of Yahoo's wasted decade: Bradley Horowitz left for Google, Flickr's Stewart Butterfield founded Slack, and Chad Dickerson went on to run Etsy. The corporate-incubator format itself outlived Yahoo — the structure Google X, Area 120 and Microsoft Garage later adopted with more patience and funding. Brickhouse's two years remain the standard case study in what happens when a struggling company buys innovation by the room.