A Lean Startup Tryout

At a recent warm up event for the Dublin Web Summit, Eric Ries spoke to a large group of people about his experience writing software that never saw the light of day as a completed product. There are various reasons why this has happened, but mainly it comes down to not building something that people actually want.

We are all aware that writing software is a complex and expensive thing to do, so such abandoned projects represent a tremendous waste of time and effort that could be better spent. The software business is littered with failed startups, that built products that nobody wanted, or were not willing to pay enough for to justify the investment.

So he has taken this experience and created a new methodology for evaluating ideas, called the Lean Startup. Central to this methodology is the concept of carefully designing experiments to test each assumption being made in creating new product. In addition, these experiments should be executed quickly and the results evaluated in a series of iterations. This is done by building what's called a Minimal Viable Product at each iteration. Such a product has the minimum amount of features to test a particular aspect of an idea. In this case, the definition of product is quite broad and may or may not involve an actual software product. (For more information on the Lean Startup methodology, check out this slide presentation or have a look at Eric's blog.)

After this talk, I wondered if it was indeed possible to test my ideas in this way. As someone who learned to code as a teenager, my computer is littered with such incomplete projects; fragments of code representing ideas that briefly blossomed, only to wither away when my attention moved on to something else. I always wonder though what would have happened if I actually brought some of these ideas to a reasonable conclusion.

So I decided to try it out and see where it led.

There is one idea in particular that has been bugging me for about a year. I had spent weekends writing code to solve a problem that I have personally, but without really thinking if this was the right thing to do. For the moment, I will call this, the Problem (with a capital 'P').

Applying the Lean Startup principles to this Problem, I created my first Minimal Viable Product - a survey. Before wasting any more time writing code, I needed to find out if other people had this Problem too and if I created a solution, would they use it.

I came across this blog post by Lindsey Harper about testing ideas using a survey managed with Amazon's Mechanical Turk. This is a service where you can pay a small amount of money to people online to complete a simple task. So here was a simple and effective way to get some answers to my questions.

Firstly, I opened up the survey to a small pilot group of 30 to test the wording of the questions. Within two hours, I had my results. Piloting the survey like this was a very useful step, as it turned out that one question in particular was misinterpreted by almost all of the respondents and required redrafting.

Following this, I launched the survey to a much wider group of users and when the results are in, I'll post them, so watch this space.

If you have tried the Lean Startup principles and have any comments on the approach, I'd love to hear them!