Songza worked on its music streaming product for over a year before it struck gold. Since launching its “Concierge” service in March, the service has blown up. It wasn’t a direct line to success–I spoke with Songza co-founder Elias Roman about what he learned building Songza and his first startup, AmieStreet.

But first, the success story. Today Songza unveiled some promising signs of momentum:

  • The company raised a $1.5 million convertible note from a long list of notable investors (below).
  • Since June, two million new users have joined Songza and that many users were active it in the last month.
  • The app launched in Canada last week and is already the number one iOS app in the country.

The company hasn’t let on how it will monetize, except to say it will not be with advertising. Judging by the list of investors in Songza’s latest round, I would bet it has something to do with commerce and data. Possibly celebrities.

Songa’s funding was led by Amazon, which acquired Songza’s founders’ first company, AmieStreet, in 2011. Prior investors Deep Fork Capital and Metamorphic Ventures participated, as did William Morris Endeavor, 1-800-Flowers (which is doing venture investments now?), Troy Carter, subscription commerce guru Brian Lee, Jay Cohen, NBA star Baron Davis, Geoff Judge and John McAteer.

The biggest lesson Roman says he learned with Songza was to ensure his product was radically different but mercifully simple. When Songza launched in 2010, it was like a pre-Turntable.fm. A function called Social Radio allowed users to be DJs for their friends. “We got pretty excited about this – after all, it was super buzzword friendly: social, crowd-sourced, real-time,” Roman says.

The problem was, it wasn’t terribly differentiated from anything out there. And then a couple of important data points became apparent. For one, there were more features than users wanted. And two, the quality of the crowdsourced stations was really low. “We had a curation problem,” he says.

Rather than invest in fixing that problem, the two data points taken together suggested a more exciting direction for the company.

The company started with the small objective of recruiting 25 experts to make playlists. They did that to see how the product felt with real content in the system. The expert playlists became so popular that Songza quickly made that the company’s mission: Build the greatest collection of expertly-curated playlists the world has ever seen.

Once Songza launched Concierge, the service took off (for good reason; it’s excellent).

Now, when I go to Songza.com, I see a prompt: “It’s Tuesday afternoon.” Correct. Songza asks what kind of playlist would be appropriate for my situation. My Tuesday afternoon options are “Easing the Tension,” Summertime,” “Taking the Day Off,” “Work or Study,” and “A Weekday Dance Party.” I chose “Work or Study,” and then within that category, “Upbeat Indie.” The concierge gets as situationally specific as driving to the beach, coming down from a night out, or cooking Italian food.

It is by design, mercifully simple. Once the company realized it had something good, the team took what Roman calls “aggressively-focused positions.” They stopped thinking about and accommodating for edge cases. Any feature suggestion that might appeal only to early adopters was immediately killed.

Songza could have easily been just another random streaming service. Instead, the company has succeeded at its very focused goal of building the greatest collection of playlists. And that’s proven today by its rapid adoption.