Don’t build your business under someone else’s roof

A few weeks ago, Twitter announced that its API would no longer be freely available. Suddenly, dozens of third-party applications were unpleasantly reminded that they don’t fully control their business. You may argue that introducing a paywall is not that impactful, and in some cases, you’d be right. Many of these applications can indeed pay the bill. But the point here is not money, it’s control. If you have a product built around an external service, the rules that dictate how your product works can change at any given moment, and you have no voice whatsoever.

This is not a jab on Twitter (although they do deserve it). They’re just providing the more recent example of what can happen when you’re living under someone else’s roof. If you look a little further, you’ll find a few others. Twitter itself banned third-party clients not long ago. The same third-party clients that made Twitter what it is today. Products like Tweetbot, which have served people for more than a decade, were gone overnight.

There’s no other way to put it, if you value control, you need to build your business so that its value proposition is not tied to something owned by another entity. The way I see it, it goes way beyond not building applications whose goals are directly related to other services or products. Any closed platform fits the danger zone I’m describing here. For example, mobile applications. As much as you’re free to create whatever mobile application you want, you don’t control its distribution. Apple and Google do. You must follow their rules, and hope for the best. And even if you do follow their rules, your application can still face some problems.

However, it’s also important to draw a line. 100% of control is an impossible and impractical dream. You probably don’t want to host your applications on your own servers or host your own email service. It’s too much trouble. You need to rely on other services and companies on some level. But this level should be the same as the examples I’ve just given: basic services. Things that don’t lock you to a specific provider. If your cloud provider decides to randomly delete your web applications, you can host them somewhere else. If your payment gateway decides to ban your account for whatever vague reason, you can migrate to another one. I’m not saying it’d be a minor inconvenience, but it wouldn’t take you out of the business. The same can’t be said if what gives meaning to your product is someone else’s product.

I know what I’m advocating here is not always possible or desirable. Sometimes you have no choice. If your product can only exist as an enhancement to another product, then the only alternative would be not to build it. And I don’t have the nerve to say it’s better not to build anything at all. Or maybe you believe that the benefits outweigh the risks, and you want to try anyway. In both cases, it’s important that at least you’re aware you’re gambling, and your whole operation can go kaput. Unceremoniously and without previous notice. Hope for the best, but also be prepared for the worst.