Reverse startup’s MVP

My guess is that most of all start-up entrepreneurs are wrong!

Just the way I was wrong when I started my previous companies. But this year I’m pretty confident to say that in my new project, Dixibox, we made things perfectly right. Basically, we started from the opposite end.

Let me elaborate that idea…

The old way

Imagine an entrepreneur who wants to create a startup that provides some kind of solution for finding restaurants in your city. Most entrepreneurs would start developing some demo app that shows a map and places some pins all over it. They will focus first on creating the algorithms for locating those restaurants on the map based on a list of restaurants stored in a database or a json file. Once that works, they will start thinking on making the app nice, all the menus, all the backend to manage user accounts, etc. Basically they leave the boring part for the end.

I believe that’s the natural way. And that happened to me when creating Mobincube, a recognised SaaS platform for non-developers to build Android and iOS apps. Back in 2007, I started at home developing some algorithms to turn a XML file into a mobile app. I was supercool for that time, quite geeky. Then we decided that it could turn into a cool product that a lot of people would need. So we started developing a web interface so people could edit that XML file just by dragging and dropping. After that, we implemented the signup forms, so Internet users could register. We made them pay by wire transfer. Then we integrated with some payments platform, so our users could pay directly from our website, etc.

The result was not bad. Actually it was pretty good in terms of product. But from a technical point of view, we built a dinosaur full of patches. The database that we built at the beginning for the app building interface conditioned all the following developments, so we’ve spent excessive amount of money and resources trying to make everything fit the early pieces of our software. I could say that we used more than 4 people during at least 2 years to build and maintain the part that allows users to pay their monthly subscription.

Developing the MVP from the end

After that experience, the story of Dixibox has been totally different and we’re superhappy with the result. On July 4th, 2018, I found my cofounders, including 2 great programmers, and we started the company. Six months later, on January 8th, 2019, we’ve released our MVP.

The funny part is that we didn’t start developing the product itself until December. We’ve left the core business development for the end.

Dixibox is some kind of voice mail platform. You can see it as a personal (or corporate) voice inbox on the cloud, that allows you to create, store and share voice notes.

Before December, we didn’t even try to develop algorithms to record audio! From July to November, we’ve been focused on developing the boring part: the skeleton of our website, the user management system, the subscriptions management, invoicing, etc. We spent months doing and redoing things. Long meetings trying to solve all possible user scenarios, etc. We always had that feeling that we were not making any progress, since we didn’t have anything that had to do with the voice mailing thing.

We didn’t want to start with the product until all the boring part was finished and rock solid. Finally, we started developing the product in December and it’s taken around 4 weeks to finish our MVP.

Why is this new way better?

Yeah, I know that we could have reached the same MPV in 6 months doing things the old way. But, why are we so happy with this new way of developing our MVP?

Pretty simple: we’ve developed a super robust skeleton that is totally independent from the product, and we’ve used modern web and infrastructure technologies. Just to name some: kubernetes, react, cypress, continuos integration, etc.

Dixibox is an MVP right now. But… what if Dixibox doesn’t succeed as a product? We would need to pivot and create a new product. But next time, we won’t need 6 months, since all the boring part is done and it can be 100% reusable. With the old way, the boring part would be really mixed with the product part, so it wouldn’t be reusable.

We could even sell the skeleton to other entrepreneurs who want to start a SaaS company, so they could save months of development and tens of thosands of dollars. In fact, you should talk to me in case you’re planning to start your own start-up, we could provide the boring part ;)

In fact, Dixibox is not my last project. We plan to create some more startups in the next years, and we will definitely save some bucks on those!