I’m working on Gush, a startup around photos.

One thing that I’m doing different this time, is that we measure *everything*. Literally everything. And I noticed that that creates a mindset.

Whenever I am working on a new feature now, I think, while sketching, how am I going to measure this. What is success going to look like in numbers?

And from seeing some numbers already, I also know that small decisions make big differences. If I am working on a feature and I know only 5% of users are ever going to use it, is it worth even working on? If I am putting a feature behind a click, and I know nobody will ever click it, is it even working on.

I don’t think the measuring mindset, by itself, leads to success. You risk optimizing into a local maximum. But it does lead to better product decisions.

Giff Constable writes down almost exactly my own thoughts on the lean startup movement.

In particular, it’s almost funny how a lot of the research techniques are old hat to us in the UX world. On the other hand, UX has never been strong at doing the type of “but do they even want this?” type of early phase research lean is focusing on, and I think that, plus a strong theory around “everything is a hypothesis, and your product is there to test that” is what I think Ries & Co are adding. A great, great contribution for sure. Much better than what the UX folks with their “let’s do some ethnography early on” talk.

UX never figured out a great way to do early hypothesis testing. They didn’t create a language around it. Lean startups did.

I think the underlying reason for this, and for much of UX’s frustrations and limitations, is that the UX role in a company usually isn’t the decision making role. Whereas, by focusing on startups, lean startups speaks directly to the people making the final decisions. UX people have gotten them stuck in consulting roles. Too bad.

Enough unprocessed thoughts on this. More later.