40 engineers is a "skeleton crew"? 40 engineers? There are tons of websites/products out there doing a lot more with a lot less.
And I love pandora, but this product management technique isn't exactly revolutionary. It is just a pretty standard top down system, albeit only planning one quarter at a time. This isn't pushing the boundaries like Valve or Yammer.
The number of users isn't actually what makes it impressive. At Napster, we had 80 million users at a time where we had 1 backend engineer, 3 web developers, 3 windows developers and 3 mac developers.
But 40 engineers to build apps for web, iOS, android, blackberry, windows phone, a thousand consumer devices, 100 types of cars, etc. is impressive assuming they roll out new features regularly. Even if some of it was done third-party, just the managing new integrations, ensuring feature releases are backwards compatible and testing new releases is still a feat.
I agree the project management style isn't anything special though. Plenty of people use something like scrum with dollar values attached to features.
Don't know about Yammer, but Valve has a very flat structure where engineers work directly on projects that they are interested in (and presumably that they think will bring in revenue). Quote from their (definitely must-read) employee handbook:
"We’ve heard that other companies have people allocate a percentage of their time to self-directed projects. At Valve, that percentage is 100."
It's hard to compare across different types of products, but IIRC, Instagram had only 3 engineers well into the millions of users (which I always thought was pretty impressive).
Interestingly, I worked there when there were 4 engineers. They had most of the product worked out back then. The Pandora website the last time I checked is not terribly functionally different from the internal Savage Beast website from 2001. Pick a song, and it generates a playlist based on that song. The barriers were legal and monetary, more than technical.
Were the ideas for new features and improvements all internally generated? Did they not have some method for actual users to suggest and vote on features?
And I love pandora, but this product management technique isn't exactly revolutionary. It is just a pretty standard top down system, albeit only planning one quarter at a time. This isn't pushing the boundaries like Valve or Yammer.