Minimum Viable Experimentation

Those who work in the digital and agile world should be pretty familiar by now with the implementation of a Minimum Viable Product (MVP). This is the creation of a working product that doesn’t have to meet all requirements, but allows further testing, feedback & iterative improvements. It is an approach pretty well understood and used across the industry and one that should lead to better products sooner.

So perhaps we need to use this approach, not just in the creation of the initial product, but in the way we release further functionality & features to our products? This would mean focusing less on the usefulness what each new piece of functionality provides (in economic terms the 'utility'), but basing each successive development on what it tells us about the product's overall ability to meet the wider strategic objective?

In other words, rather than add new stuff that simply compliments the overall richness of the experience... shouldn't each new tangible delivery be based upon a hypothesis? And in-turn, shouldn't this hypothesis be derived from insight that is focused on improving the user's needs or outcomes?

For example... if your project aim (which I assume directly linked to your strategic objective) is "to have a better online sign-up process for a new credit card", then each successive sprint or release from the initial product launch should be delivered to address this aim. However you shouldn't just assume that this is the case. 

Firstly make sure that each time you plan your deliverables you are actually answering a question, such as:

"How can we stop [a specific type of customer] exiting the online form before the end of the process?"

Secondly develop a hypothesis that can be tested in a small experiment. Such as:

"We believe that by adding [1: a specific feature] at [2: a specific point] we will create [3: an expected behaviour] by the user and therefore they will reach [4: an outcome] that will improve [5: a goal]."

Where in our new credit card sign-up example this could be:

1	A specific feature	A reassuring statement about financial approval
2	A specific point	The 4th of 5 pages in the process, where the most users drop out
3	An expected behaviour	The user is reassured that they could be approved easily
4	An outcome	The user moves to the 5th (and final) page of the process
5	A goal	Form conversion improves

It is worth stressing the point that these experiments don't have to be huge or complex, and in some cases making changes to a piece of content or image may be sufficient. They just have to be enough to prove or disprove your experiment's hypotheses…. a minimum viable experiment and nothing more.