Draw Something: A Lesson in Scalability

3 min read
Aug 23, 2017

Unless you have been living under a rock, or indeed outside of a suitable 3G or WiFi network, you will have heard of Draw Something. This ubiquitous little app from a relatively unknown publisher recently shot to the top of the charts before being bought a mere two months later by Zynga for a sum of $200 million. So what is it that made this app so popular and what effect did this explosion of users have?

The Ingenious

Let’s first look at what made this such a pull for people to use. The key to Draw Something is in its social element, you can play against random Facebook people but the real fun comes when you play against people you know. When you invite a friend to play, there is already a level of brand confidence, after-all, you trust your friends and if they use it then it must be good! This approach works very well and is the key reason people put social sharing features in their iphone apps, it’s the easiest form of marketing there is. But Draw Something took it one step further, an ingenious if a little annoying step.

Each game on Draw Something consists of a watch-guess-draw combination. You first watch the other player guess your last move, then you guess the other players drawing and finally you draw your own for them to guess. The clever part here is that once you enter this cycle with a player you can’t get out until you complete the three steps, this results in a game ALWAYS being in play. Clever huh.

These clever approaches to user engagement meant that more and more people were invited, and subsequently locked-in, to Draw Something and is one of the major reasons it shot to the top of the charts. But with such a meteoric rise to fame comes the inherent issues of scalability, and sadly Draw Something were no exception.

The Not-so-Ingenious

When developing iPhone apps, or indeed iPad apps, where server-side elements are required, deciding on the requirements is a tricky thing. The optimist would opt for a rack of ten or more servers, redundant power supplies, banks and banks of storage and the best processors and memory available. A costly enterprise. The pessimist would attempt to run the server on a home-grown PC in their bedroom with a mediocre internet connection. A positively cheap alternative. The simple fact is that it is very hard to know up front how successful an app will be. If it were that simple we would avoid the apps which we knew would fail! Getting it wrong can be costly, as Draw Somethings creators soon found out.

Underachievers Can Leave Now

When Draw Something exploded, the back-end systems were simply not able to cope with the demand. People who had just downloaded the app were faced with timeouts, crashes and a poor user experience. In this case the tech guys were diligent enough to rectify the issue within a day or so, but for others the issues are left untouched and the customers perception of the app, and conversely the ratings, drop considerably; and once they drop it is very difficult to regain them. Underestimating the potential success of your app can be disastrous, so think carefully before installing that server your the bedroom!

All Better Now?

So now that Zynga have bought Draw Something this situation must be better right? Well, no, not quite. You see with any tech buy-out comes a hash of teams, approaches, technology and ideas, and this mash-up was no exception. 

Unless you have been living under a rock, or indeed outside of a suitable 3G or WiFi network, you will have heard of Draw Something. Click To Tweet

With Zynga being stalwort Facebook developers, the extent to which the app was integrated increased exponentially, and increased very quickly. The rush to get this integration work done has caused many users to become frustrated with the longer process to play, the crashes which cause them to lose points and the more in-your-face tie-in with Facebook. The user base for Draw Something has dropped significantly since the buy-out, so perhaps it’s time to roll-back some of those changes…

Get Email Notifications