My business partner Greg and I always joke about how one of the biggest differences between college and the real world – specifically running a business – is that in school you almost always have all of the information that you need to make a decision or solve a problem, whereas when you’re running a small business you rarely do. The funny thing is that he’s basing this on his MBA classes and I’m basing it on my engineering classes, which on the surface one would think couldn’t be more different.
This problem is especially evident when you’re building something new and you have literally nothing to work with. One of the best ways to overcome it and make a decision that’s generally mostly right on the first try is to study what companies have done in similar situations that’s worked. And not just study what they did, but why they did it and why it worked. Essentially, use their data. Or maybe more accurately, use the conclusions drawn from their data to draw your own similar conclusions. Your decision probably won’t be perfect, but it will be good enough to get you in the game.
This recently came up with LockerPulse. One of the first decisions that we made when working on the new LockerPulse was to separate it in to two distinct sites: the application that logged-in users see, and a “marketing site” for registrations, logins, help, informational pages, our blog, etc. Why? Well we noticed that that’s what most other successful web apps do, and with good reason. It’s too complex to try to do everything in a single template as we do now. You end up compromising the quality of the app as well as the effectiveness of the signup process.
A few weeks ago we sat down to start planning out the marketing site. We’ve never built a marketing site before. The #1 goal by far is to convert new visitors in to registered users at a much higher rate than we’ve done in the past. But what we know about converting a sale on DI doesn’t really apply, so we didn’t have much to work with beyond our own opinions.
Instead of diving right in to a design, we took a step back. We made a list of roughly twenty similar web apps (Mint and Evernote are two examples). Most of them have been around for a while, are extremely popular, extremely well funded, are built by really talented teams, and – most importantly – have signed up hundreds of thousands of users. If you think about that for a second then you realize that they’ve likely gone through dozens (maybe hundreds) of iterations before arriving at their current marketing site. Every detail about the design and functionality is the way it is now because it was supported by their data.
When you look at twenty of these sites in a row you start to pick up patterns. Patterns in the design, patterns in the registration process, patterns in the ad copy, patterns in the call-to-action, patterns in the link structure, and much more. At that point we could sit down and mock up our site with quite a bit of certainty that we are going to do is going to work. Or at least not completely suck. Which is really the best you can hope for when you have nothing to work with.