Value vs Waste: Customer Validation

I have come to learn, working for a startup for the past few months, the importance of Customer Validation. Granted, I have been immersing myself in one of my favourite books, The Lean Startup by Eric Ries, but correlating the chapter’s philosophies with what’s happening on the ground, is understanding and evaluating what is value versus waste? man-162604_640

We could spend a lot of time creating our MVP, or Minimum Valuable Product, deciding what features to include, what features not to include, but without actually understanding what features are needed in order to gain actual feedback that we can measure our decisions against, we might as well be spending time on wasteful directions.

The author describes his former company’s decision to support multiple networks in it’s initial MVP, and eventually found out that their assumptions about what customers want was flawed. Could he have determined that prior to going down the track of spending all that extra effort? Sure.

..Learning is the essential unit of progress for startups (Ries, 2011).

The point is, if we are not going to learn anything about what customers want from a feature, then we can defer it, or eliminate it. This is the essence of validating, gauging positive improvements in core metrics (referred to as Lean Analytics) you look at what customers really want. You do this through experiments, not through questionnaires, you provide a decent-fidelity prototype for customers to play with, and observe. You do an A|B Testing methodology, to see from different versions which one gets the most positive responses.

What is considered vanity or actually useful metrics or stats is outside the scope of this post, and I will certainly look at including it in a future post, but it’s important to note that numbers and downloads and registrations are not important to startups. What’s important is correlating what feature you include is a positive attribution, rather than correlating increase in number of downloads based on each update, assuming that your features must therefore be what they want.