Why Google Bought Songza

songza google colors cropIn a move that’s been rumored for several weeks now, Google announced yesterday that they’re buying the personalizable multichannel webcaster Songza, for a price that was rumored yesterday to be in the $15 million range but which the New York Times today estimated to be $39 million.

Google now will have three different music streaming services: Google Play Music All Access, YouTube’s upcoming streaming service, and Songza (unless it is absorbed into the former).

KurtHanson-modSongza’s unique angle is a front-end on its homepage called “Music Concierge.” Songza’s homepage allows you to set the current day of the week, time of day, and activity you’re involved in (from a half-dozen choices like “Washing my dog”) and then offers you a series of four to six choices of styles of music (from a handful of choices for each activity like “Peppy Indie Pop/Rock” or “Summertime Oldies”). Once you’ve selected a style of music, Songza gives you a choice of three specific playlists within that style to listen to.

Each playlist is about 40 to 100 songs long, which provides the listener with about two to five hours of music listening before songs start repeating.

According to CEO Elias Roman’s timely interview with RAIN’s Brad Hill on Monday, the company has about 60 freelancers (along with a reported 40 full-time employees) curating playlists from their homes, with the benefit of a dashboard that provides them with historical user ratings (e.g., thumbs up, thumbs down, and skips) for each song.

The company claims to have 5.5 million monthly users (as compared to a similar figure of 75 million from industry leader Pandora), although this is a self-reported numbers as Songza does not participate in the webcasting industry’s primary ratings service, Triton Digital’s Webcast Metrics. (Or, if it is a Triton client, it doesn’t allow its numbers to be published in Triton’s monthly “Top 20” press releases.)

In its reporting this morning, the Wall Street Journal’s Rolfe Winkler wrote, “Songza had raised about $7 million from investors including Lerer Ventures, Metamorphic Ventures and Deep Fork Capital,” but in the New York Times, reporter Ben Sisario noted that “counting investments made in Amie Street” — the co-founders’ predecessor company — “Songza has raised a total of $12 million from investors including Amazon, the William Morris Endeavor talent agency and the music managers Troy Carter and Scooter Braun.” It is unclear how much of the equity is still owned by its co-founders, including the aforementioned Roman and Eric Davich. (According to company more, the firm’s initial co-founders met a few years ago while at Brown University.)

The purchase seems like an odd one for Google, given its historical preference for algorithmically-created products (e.g., Google News).

The problem, however, is that music is difficult to program on a purely algorithmic basis: An intense Seals & Crofts song, for example, may sound acoustically similar to a mellow Led Zeppelin song, but an experienced music programmer (i.e., a human being) would be able to discern that they’re culturally very dissimilar.

Pandora’s “Music Genome Project,” in fact, categorizes songs based on their acoustic characteristics but is completely unaware, as founder Tim Westergren used to reveal in his cross-country series of “town hall meetings,” of the songs’ lyrics.

While $39 million may seem like a high price for Google to pay for an attractive front-end and a couple of thousand short playlists, the truth is that building a full-featured personalizable multichannel online radio property (including acquiring music tracks, programming the channels, building databases containing song ratings and skip histories, and customizing separate streams of programming for each listener) is a long, difficult, and expensive process. Apple learned that with its disappointing launch of iTunes Radio last fall.

Every webcaster that has built such a full-featured platform (including Pandora, Slacker, AccuRadio, 8tracks, and Songza in the U.S. and Aupeo in Europe) has spent years of time and millions of dollars doing so.

But it’s clear for quite a few years now that music radio listening is moving from FM delivery to Internet delivery — thanks to consumer benefits like lower spot loads, attractive visuals, long-tail formats, deeper playlists, and personalization — so this may be a smart investment in Google’s future.

Kurt Hanson