In his book The Year Without Pants, about the year he spent managing a team at Automattic, Scott Berkun outlines the approach that the firm uses to build new features. It’s a pretty good method for building an experimental culture. First, he says this about experiments and innovation:
The fundamental mistake companies that talk about innovation make is keeping barriers to entry high. They make it hard to even try out ideas, blind to how much experimentation you need to sort the good ideas from the bad. I’ve visited companies that use big meetings, with far too many cooks in the room, to rank ideas based on one-sentence descriptions. It was madness. While few at Automattic had ever experienced this, their process was on the other end of the spectrum: they had faith in the future. If a feature launched and had value, the missing things could be addressed later. And if after launch it didn’t have value, few would use it and it could be killed (although I’d learn later that killing of experiments was far too rare).
Here are the seven steps that they use at Automattic to build new features for WordPress.com:
- Pick a problem. Right off the bat this is a different approach than many organisations use. Often people are try to push a cool idea that they’ve had, hoping that people will like it. Starting with a problem is more of a design-thinking approach – it gets you thinking about the customer’s perspective right from the start.
- Write a launch announcement and a support page. Not everyone will take this specific step, but the general principle is a good one. Berkun says:
The point is that if you can’t imagine a compellingly simple explanation for customers, then you don’t really understand why the feature is worth building. Writing the announcement first is a forcing function. You’re forced to question if your idea is more exciting for you as the maker than it will be for your customer. If it is, rethink the idea or pick a different one.
- Consider what data will tell you it works. Figuring out what success looks like will make it easier to kill ideas that don’t work, freeing up resources to try other experiments.
- Get to work. Build it! Focus on the goal you outline in step two, but be willing to pivot the idea once your idea starts to become real.
- Launch. At Automattic, it sounds like they’re about halfway in between launching a minimum viable product and one with full features. They never launch with everything in place – they iterate based on the data that they’re measuring from step 3.
- Learn. Learning is what distinguishes experiments from mistakes.
- Repeat. It’s always a loop, not a line.
This approach raises some interesting points. One is the idea of the forcing function – building in a point that requires you to define your objectives. In part, this is where innovation intersects with strategy. More importantly, it is good practice to include steps that force you to think like your customers – a point made well by Ralph-Christian Ohr here. This is basically another version of Simon Sinek’s Start With Why idea.
The second big idea is the focus on learning. If you try an idea and it doesn’t work, and you don’t learn anything from this, then it really is a failure. None of us have enough spare resources to afford this. Nevertheless, to innovate we have to try out a fair number of ideas that end up not working as we expected. This is only feasible if we structure things so that we learn from our experiments.
The third point is that very often our really big ideas start out pretty small. One of the projects that Berkun’s team worked on ended up being JetPack. This is the biggest feature that WordPress has added to their interface, I think (and it’s awesome). But it didn’t start big – it started out as an effort to coordinate the services provided to customers with self-hosted blogs (like this one) and to customers with their blogs hosted on WordPress.com. It was only as they built it that it grew as big as it ultimately became.
The seven steps clearly come from an online business. Nevertheless, with a bit of tweaking you could use this approach in nearly every type of organisation. And since experimentation is such an important skill to build, it is worthwhile to do the tweaking.
So find some things to work on, and get to work. Start small, and who knows where you’ll end up?