20% Time and the New Google

March 2013

In late 2011, I wrote a piece arguing that Google operated as an applied research lab, with employees encouraged to pursue interesting ideas as ends-in-themselves. Over the past decade, Google’s search engine has basically functioned as a cash cow, allowing the company to develop two freely-distributed operating systems, self-driving cars, augmented reality glasses, and a myriad of other projects, many with no clear revenue stream.

I think this is changing. Stephen Levy’s In The Plex was published in early 2011, and I think that book serves as a comprehensive history of ‘Old Google’. Notably, that book ends with Google planning to launch their social offering, Google+, and prefaces Google transitioning to become a different kind of company entirely.

Google+ is not an Old Google product. Old Google products were defined by compelling features that pushed a particular market forward: a search engine radically better at indexing the web than competitors, a webmail service that provided an order of magnitude more storage than the incumbents, an online mapping solution that defined the market. More crucially, Old Google products often had no real strategic connection to one another: it often seemed as though Google engineers would simply begin by exploring an interesting space, and keep experimenting within that space, testing its limits, forming a product (sometimes in public) as a byproduct of that experimentation.

Google+, on the other hand, was almost certainly spearheaded not by engineers working from the ground-up to make something cool, but from the top-down by executives convinced that Google needed to make its own version of Facebook. That’s why Google+ feels like a refined but incredibly boring product: it’s not something that any engineers at Google would have decided to work on if really given the choice. It’s not a 20% Time project.

The New Google, as Marco Arment suggests, is more strategic, colder, and more focussed. That’s kind of a good thing. It means that the products and services Google produce over the next few years will likely fit together more cohesively, and ship with a higher level of polish (remember Google’s perpetual betas, or – shudder – early Android?).

But that’s also the problem. Creating products to fit into an overarching corporate strategy is not very fun, particularly if you feel as though you have little power to shape that strategy. I think, as power has slowly trickled up to the corporate level, Google will begin to find it extremely difficult to recruit and retain engineering talent. Apple, at least, have always had a strict corporate hierarchy, but Google are still masquerading as though their corporate structure is radically ‘flat’.

A flat structure gets you great, weird products, produced by passionate engineers, and a hierarchical structure gets you a cohesive set of products that work flawlessly. The issue is that I think Google want it both ways, which makes no sense to me. Google have always made a big deal of their 20% Time philosophy, which involves giving an engineer one day a week in order to play around with self-directed projects. But here’s the rub: how can you empower engineers to continue creating weird, cool products in their 20% Time, while at the same time killing existing weird, cool products in a bid to ‘streamline’? Either you tell your engineers they’re free to create cool stuff, or tell them their job is to follow orders from the top in order to create refined, vertically-integrated experiences. 20% Time doesn’t seem to fit with New Google’s obsession with focussing on fewer products in order to “make for a better user experience”.

The real problem for New Google is not that top-down, focussed product development is a bad idea, but that it runs counter to the very strong company culture fostered by Old Google. If Google pretend that they’re still all about empowering engineers to develop awesome, slightly rough products from the bottom-up, they’ll piss more and more of those engineers off once they find that’s not really the case.