The Wonder of Emergent Behaviour

I was watching my 2 year old son play this morning. He was playing with a Fisher Price cash register. It’s a fairly complex tool for little kids, but he loves it. The general workflow is that you put a coin in a holder in the top where it stays until you hit the appropriate lever. The slot opens allowing the coin to fall below into another holder. You then push “sale” and the coins drop from the 2nd tier into the cash tray at the bottom. Alternatively, you can push refund and watch the coins come out the side. When you’re all done, you turn the handle and the draw opens with a “ding”. Rinse and repeat.

What amused me this morning was that rather than pushing the “sale” button to make the coins drop from the 2nd tier he was loading the top tier with another set of coins. The problem here is that when you push the lever the coins can’t fall to the 2nd tier as the previous coin is still there. Frustration ensued, requiring me to calm him down and explain (again) how it works.

But after a while something interesting happened. When he pushed the lever a second time on the jammed coin the lever become stuck and after about 3-4 seconds flung back and hit the coin with a satisfying “THWACK”. He screamed in surprise and ran away from the register as if it had bitten him. Moments later he was back intentionally doing the same thing. With each “THWACK” he’d scream and run away in feigned shock. Through not following the designated instructions, he’d discovered an entirely new way to play, one which was arguably more fun than the origin intent.

Of course, anyone with children won’t find this sort of behaviour very surprising. Kids have no pre-conceived idea of how to use a toy or how an object reacts until they play with it and see what happens. The interesting thing is to find that adults are more than capable of producing some of these startling results themselves.

We’ve been running BugHerd in beta for about 2 months now. Our original intent for the application was simply as a means for two web developers to log and communicate issues. We soon discovered the real value was between technical and non-technical users, and have began pivoting the product in that direction. In short, BugHerd allows you to log a bug as in a traditional bug tracker, but by clicking on an element in a web page and tagging the bug against that element.

The amazing thing for me is that every time we demo BugHerd we see and hear new ideas about what BugHerd means to that person. What was supposed to be just a simple bug tracker turns out to be of significant value in ways that we’d never imagined.

One classic example we had from a gentleman from Lenovo we did a demo for at the Web2.0 expo. His jaw dropped when we showed him what BugHerd could do, but it wasn’t because it’s a great bug tracker. He had another use case so far out of left field that were simply speechless. It turns out that the standard way for a company to have their web site translated to other languages is to send a spreadsheet or XML file full of text to a native speaker for translation. The translator often has no sense of context, and for short passages of text, calls to action or titles can easily be mis-translated.

Until now this has been a problem that has been near impossible to solve. However, he saw immense value in getting translators to use BugHerd to go through the site tagging specific text elements and providing the translation. With context comes understanding, something we had intended for bug tracking. Through this emergent behaviour we discovered a whole new use case that we’d never remotely considered.

Another classic example actually came from one of our investors. He emailed me early one morning after having a dream about BugHerd. He exclaimed “EUREKA!” as a he realised that a user could use our tool to log bugs on ANY website in the world with the use of a simple bookmarklet. We could then store the bug and let the user log the issue directly to the company concerned. The website owner could then come to BugHerd and see exactly where the issue occurred all without having to have any further contact with the logger. The entire world could become your testers if only you could give them the tools to do it…and as it turns out we can.

Of course the danger to emergent behaviour is maintaining focus on your vision. It’s all well and good to get excited about new ideas and new possibilities, but at the end of the day you need to build what it is you think is the “best” use of your application. Fisher Price clearly shouldn’t change their cash register to work the way young Angus decided it should work, and nor will BugHerd become a translation tool any time soon. These moments of brilliant observation are often distractions, but every now and then they highlight a lateral approach to a problem previously unsolved. Sometimes letting users experiment provides the greatest insight into the real value of your application.

 

About Alan - Alan Downie has been working in web for over 20 years. Alan was the founder of Fivesecondtest, UsabilityHub and BugHerd. He participated in the Startmate (Australia) and 500 Startups (US) accelerator programs, before becoming a mentor at partner at Startmate. He recently launched Splitrock Studio, a startup studio based in Melbourne, Australia.

Comments are closed here.