I think he's addressing the riskiness of your first DS hires. PhDs have a much higher likelihood of possessing these skills than non-PhDs. In his defense:
(1) "[T]o be competitive" means a lot in Pandora's case. They pioneered music recommendations, so they needed autonomous scientists with strong research capabilities. The vast majority of companies do not fit this profile.
(2) Earlier on he states
"Instead of setting your sights only on PhDs and research scientists, you need jacks-of-all-trades who tend to be more interested in practical applications than theory."
He's guilty of mixing absolutes: you NEED these types but hiring PhDs is "crucial." Why not hire PhDs with 10+ years of coding and product experience at slightly below market + equity while you're at it?
Your point about experience beyond academia is crucial. Recent PhDs with little or no commercial experience are a really risky bet as a first hire since they'll need to shed the academic cruft they carry before becoming productive.
I'm astounded that they only got a "data scientist" 3 years ago. Wonder who/what controlled their recommendation engine before that?
All-in-all that was a pretty well balanced article. I like the bit about PhDs - don't need 'em, but useful since their whole schtick is solving hard problems independently.
Recommendations were likely produced by applying collaborative filtering to their expert-extracted features (the whole "Music Genome" thing). Collaborative filtering is easy to implement, and the massive feature extraction investment likely provided a rich dataset from which they could recommend songs.
