Experimentation: Why it requires rigour

If more and more companies are responding to the siren song of experimentation, doing it right may well become an important building block in gaining competitive advantage.

By Suvi Nenonen and Kaj Storbacka. 

There is a lot of ado about experimentation these days. As our operating environment is becoming increasingly unpredictable – be it because of digitalisation, globalisation or changing consumer preferences – it doesn’t really make sense to waste your time and energy on Soviet-style five-year plans that are obsolete from the day that they come from the printers.

Unfortunately, all too often talk about experimentation evokes a (false) sense of relief: “Phew, thank God we can use experimentation to solve this problem – that’s so much easier than following the formal strategic planning / product launch process!” Although popular anecdotes might suggest otherwise experimentation is, however, not a synonym for intuitive action or number eight wire – nor is it a licence to stop thinking. Ironically, clever experimentation requires considerable rigour.

Articulate your assumptions
The first step of successful experimentation is to articulate the main assumptions that you have about the problem at hand. Let’s imagine that we are thinking about launching a new app that helps farmers to optimise their water usage. In this case we might have assumptions about the number of farmers with water-related challenges, their willingness to take on new apps, and also about our own ability to market this new tool.

Outlining one’s assumptions about the future is surprisingly difficult – mainly because we are not used to it. In fact, it is quite possible to create a seemingly implementable business strategy without ever thinking about the underlying assumptions about yourself and the operating environment.

Being explicit about these assumptions is nevertheless crucial, as they are the guiding stars for clever experimentation: you experiment in order to validate – or invalidate – these assumptions.

Design your A/B testing scheme
The blueprint for clever experimentation doesn’t become any more unforgiving or ad hoc after you have identified your assumptions.

It all starts by ranking the assumptions based on how much evidence you have supporting them: the ones with little or no data behind them are the ones that must be tested first. Based on this you have to design your experiment.

Ideally, you should conduct A/B tests with prototypes in a real-life context in such a way that they generate measurable data about the assumption that you are trying to validate.

So, if we want to test our assumption about farmers’ willingness to embrace an app to optimise water usage, then we have to build two prototypes – one app and one “non-app” alternative such as desktop software – that we present to genuine farmers and measure which one evokes more favourable responses.

Be prepared to kill your darlings
And now starts the most painful part of the rigorous experimentation process: if the data shows you that your assumption is false, then you should accept that and formulate another assumption – which you will test later on with another set of A/B prototypes in a real-life context.

And killing your darlings is not something that comes easy to us humans.

The good news with the fact that experimentation is very difficult is the fact that it is very difficult. Very few companies have mastered the art and science of experimentation while the vast majority consider it as a synonym for ad hoc action.

So, if more and more companies are responding to the siren song of experimentation, doing it right may well become an important building block in gaining competitive advantage. 

Visited 4 times, 1 visit(s) today
Close Search Window