What was debated in the meeting, and where/how did folks agree on decisions? In my write-up, I gave an easy example of “just ship it” based on a successful AB test result. My point was I didn’t want people to wait to “rubber stamp” decisions like these.

What got debated was the strategy, the design and execution of AB tests, and in a few cases, the analysis of AB test results. In the personalization area, for instance, there was long-standing debate on how to establish a proof of concept for an all-new algorithmic approach called, “Quantum Theory.” We had noticed Pandora’s Music Genome approach and its ability to describe why you would like a certain song, based on the work of its hundreds of “music taggers” describing the specific attributes of a song: “You’ll like Jack Johnson because you like male acoustic vocalists who sing about surf culture.” We debated the tactic and what the MVP would be to test this approach. It took more than a year to successfully design, execute and demonstrate that the approach worked.

Another example: In the merchandising area, we debated the wisdom of unique movie finding tools (personalized previews playing on your homepage, for instance) and debated what would be a reasonable proof of concept for this theory to be proved/disproved. We eventually disproved the theory, but I’ll note some recent articles suggesting that this idea is being re-tested today — we’ll know if it was a successful test if we see previews rolling on our many Netflix devices in the near future.

As a final point, it wasn’t necessary for everyone in the room to agree on a decision. To a degree, the product leader was putting their reputation on the line and were given a relatively long leash to demonstrate their consumer insight and judgement. There was a fair amount of “disagree and commit” in an organization that was chock full of “fully formed adults.”

Thank you for your question and I hope this answer helps.

Gib