Fail Fast, Fail Often? No: Fail Small, Learn Fast

Published by Grahame Cox on

“Fail Fast, Fail Often”. Sounds exciting and ‘edgy’ doesn’t it. Don’t get me wrong, I’m all for pushing boundaries and a strong supporter of learning from failure……. but isn’t it better to do this in a way that doesn’t risk your colleagues’ future employment prospects? The good news is, there is another way.

I often come across a strong desire within organisations to miss out the ‘learn from failure’ bit altogether. To move into ‘action’. To go straight from first concepts into full scale development and onto launch. Further testing feels like procrastination, time wasting, when what is needed is speed. Of course, you may come across a ‘winner’ straight away but in the world I live in, getting to a winning product or service usually needs a bit more effort. Market research can only take us so far. In my view, real world ‘test, learn, refine’ loops should be part of every product or service development project. The fact that they aren’t is quite probably one of the key reasons for the high rates of innovation failure. The question is, how should we do this? 

Thomas Edison said  ” I have not failed, I have just found 10,000 ways that won’t work” – unfortunately I haven’t found too many companies happy to tolerate me failing that often without results. Simply encouraging teams to ‘Fail fast and fail often’ seems to be a license to thrill,  without risking any consequences. Do you remember when you were learning to ride a bike or drive a car? If you’d been sent out to the A1 for the first time without your stabilisers or for your first driving session, you may have learnt some very valuable lessons, very quickly. You may have failed fast. But you and your instructor may also no longer be here to tell the tale! So it is with innovation projects, we need to experiment, we need to fail to learn, but failure is much more likely to be tolerated and repeated, and even encouraged, if we do that in a measured way. The challenge is to run simple experiments to test out your ideas in a ‘real world’ environment, at an acceptable level of risk.

In the early days of my fmcg career I would run a regional test market, probably in the old Border or Tyne Tees TV regions (why was it always those?). After 6 to 9 months  we would then decide whether we had a winning product, or if further modifications were required before committing to a full launch. In today’s fast moving world, where the landscape can change completely in the course of a 6 month period, we need to find smaller, faster experiments that provide valuable learning to develop and improve your ideas. In the virtual world, the launch of ‘beta’ products is common practice, today sees the launch of a beta version of the latest edition of the highly successful ‘Black Ops’ computer game franchise; these ‘minimum viable products’ (MVP) are launched on a limited basis to selected end-users who, through their  real world user experiences and feedback, help develop and fine-tune the final product (a good read on this subject is ‘The Innovator’s Hypothesis’ by Michael Schrage). The challenge may be greater in the world of physical products and services but launching an MVP is still possible. 

So, next time you are developing a product or service, think about how you could run some simple, fast ‘real world’ experiments to launch a minimum viable product to test, learn and refine your idea. That way you can still learn from your failures, but you don’t ‘bet the company’ in doing so. In my experience this will payback in spades when you see those improved success rates and your launch and development budgets put to better use.