The IMVU Story According To Its VP of Engineering

I was at James Birchler’s talk, hosted by Rich Collins of the San Francisco Lean Startup Circle. James is the VP Engineering at IMVU. James’ account gave me another perspective of the IMVU story that is different from the somewhat sanitized/idealized account given by Eric Ries.

Eric talked about how IMVU adopted practices from the Toyata assembly line. If anybody made a mistake, nobody was allowed to check code into the source control. The whole team would stop in order to resolve the problem. Furthermore, root analysis was performed to ensure that the problem would not recur.

Great story.

James talked about the same thing. But he added that they don’t do it at IMVU anymore. They no longer could interrupt everyone whenever there was a build problem.

While IMVU may be a poster child of the Lean Startup movement, the reality is messier.

The IMVU Redesign

Another example is the IMVU UI redesign. The old IMVU, over a period of time, accmulated features. Each new UI component was an improvement. The IMVU client eventually became like the dashboard of a nuclear reactor. They knew it was complicated. But hey, who’s to argue with the improved metrics? The company felt that the UI was wrong, but nobody wanted to shoot the elephant in the room.

Then a group within the company started to work on a redesign. James said that the group at first worked in secret. The project would’ve been killed. Eventually the new simplified and polished design prevailed. With the flip of a switch (and presmuably sighs of relief), IMVU moved over to the new design. Revenue went up and up from there.

It was the right thing to do for the customers.

Thomas Kuhn Strikes

Yet the redesign was a non-evolutionary leap. If the normal activity of Lean Startup consists of experiments that lead to incremental, evolutionary improvements, then how do you achieve the sort of “paradigm shift” like IMVU’s redesign?

Vision.

For somebody like Steve Jobs, the vision is so strong, and right, that the company could just follow the vision, and likely arrive somewhere good. And what is vision good for? Delivering value. What IMVU is learning now is that nothing else matters. Just ask “is what I am doing delivering value for our customers?”

The lesson learned is that for a company, it isn’t enough to learn a set of techniques. Techniques are for normal daily operation. More importantly, you need a culture that helps you to relentlessly pursue your vision and deliver customer value.

You can see James’ slides, “Experimenting Your Way To Success”. The slide deck sticks too much to the party line, IMO. I liked James insight into company culture. IMVU is remarkable in its appetite for change and experimentation with its work processes.

Discuss on Hacker News

If you like my writing you should follow me on Twitter.