Consider the problem of problem ambiguity and the solution that analytics brings to the Lean Startup – “letting thousands decide what millions do”. Time to unpack that.

The problem of problem ambiguity refers to the creation of a business models or products which are intended to solve fuzzy, low-defined, problems. Yes! It really does happen!

One of the mentalities within the Y-Combinator hive mind is to simply produce a product and get it out there. Iterate. Dominate. Be capital efficient. Be lean. Ponies. Stickers. Double Rainbows.

And so, there are any number of firms out there that are generating solutions to problems that they don’t fully understand and loads of potential customers that are aware of ambiguous problems.

The behavior is explained by The Garbage Can theory that I’ve referenced before. An organization is an organized anarchy – a Garbage Can into which a stream of problems, solutions, energy and participants flow in. Solutions look for problems. Problems look for solutions. Everything’s made up and the points don’t matter anyway.

Perhaps the market itself is an organized anarchy too, and that just like March and Olsen back in the seventies, a Fortran program just might capture the complexity nicely and offer predictive value. Certainly, a glance at the #HBR Behavioral Economics craze would suggest that everything ancient is new again and that it can be done.

It might seem amazing that people produce things intended to solve a problem that doesn’t really exist. They do. It brings me around to this notion of The Lean Startup.

In the Lean Startup, the problem is iterated upon in addition to the design and code base. The Agile Methodology assumes the problem is defined. The Lean Methodology questions everything. It’s a point that is explained very well by Croll and Power.

I really like what Alistair Croll (@acroll) and Sean Power (@seanpower) have been doing. They communicate in a way that I get (209 slides with low word density), and there’s evidence of really strong thinking and genuine insight. They acknowledge complexity and deconstruct it. It’s great stuff.

“Letting Thousands Decide What Millions Do”

That quote originates from Claude C. Hopkins, who wrote that “before testing and sample statistics, advertising disasters were really quite common” (or something to that effect). Those words still cause me to smile, 80 years later. How I wish it was true.

This is where the analyst can go from being the reporter on progress towards goals, to becoming the strategic collaborator on problem definition. The rise of sophisticated analytical software enables analysts to understand the efficacy of marketing, perception, competitive sets, and usability. In effect, without analytics, the Lean Methodology and approach to problem solving collapses. The feedback loop is broken without the strategic collaborator.

Letting thousands decide ensures that an experience can be iterated upon quickly, as the user base isn’t large enough to generate institutional lock-in. There’s less risk and a better chance of actually learning enough about the problem to actually solve it, or one of its sub-variants, successfully.

In sum, the problem of problem ambiguity can be solved through analytics, in conjunction with a Lean Methodology.