Is Pandora’s Westergren a Pathological Liar?


Why are so many statistics and factoids Tim Westergren, Pandora’s founder, utters a distortion or lie? Is it because he doesn’t understand? Is it because he is misinformed? Or is it purposeful, a willful effort to distort and mislead?

Were he just talking up his service, we could let it pass. The problem is that Westergren seems to go out of his way to talk down local radio, fabricating false and misleading comparisons to denigrate local radio.

So we believe it is time to take a closer look at Westergren’s claims.

Westergren holds Pandora Listener Meet Ups around the country. At a recent Meet Up in New York he offered a long list of radio factoids. These were first reported here, and then repeated on various blogs including here, here, and elsewhere.

Some of his factoids appear absolute gibberish. Either he got his facts scrambled, or the reporters did. Anyone should have been able to realize they were meaningless. Yet the factoids were circulated without a single question or doubt raised.

During the meeting Westergren asserted that over the entire history of broadcast radio, less than 100,000 songs have ever been played. (This may be to counter the criticism that Pandora has too few songs in its library, well under a million.)

A little common sense ought to be enough to shoot this one down. First, only a foolish person would pretend to have an accurate count of how many different songs have been played from (say) 1920 when KDKA signed on.

Joel Whitburn says there have been over 40,000 Pop hits since 1955, over 20,000 Country hits  since 1944, and 20,000 R&B hits since 1942. That’s 80,000 hit songs in just three formats.

There are a dozen additional formats including Rock, Jazz, Classical, Easy Listening, Adult Standards, Big Band, Spanish, Gospel, and Religion each with their own music that add tens of thousands of songs that local radio stations play today.

Then there is the low end of the dial. Eclectic and ethnic stations at the non-profit end of the FM dial probably play 100,000 different songs by themselves.

And these are only the songs that can be heard right now. How likely is it that radio has only played 100,000 different songs ever? Zero.

Westergren likes to brag about the size of Pandora’s audience. He recently boasted that Pandora has more listeners in Philadelphia than WBEB. Now that Ando Media includes Pandora in their monthly reports, we can see whether this is true.

WBEB has a total weekly audience (cume) of over 1.6 million listeners.

Total registered users of Pandora are north of 50 million, including the inactive ones. We don’t hear much about how many people are actually using Pandora each week, because the number is considerably smaller, but let’s use the 50 million. Philadelphia is about 1.6% of the US, so that means Pandora probably has about 800,000 registered users in the area.

In other words, the number of Pandora’s registered users in Philadelphia is half of the number of people who listen to WBEB each week, active users considerably less.

WBEB’s advantage gets even greater when we consider Average Quarter Hour, a measure of average listenership. Using the same proportionality, at any given moment, eight times as many listeners are tuning to WBEB than those in Philadelphia who have ever registered on Pandora. (Read more on the calculation here.)

Were these the only misspoken comments Westergren has made about Pandora’s relative strength, one might overlook it. However, it is part of what appears to be a campaign of misinformation and distortions.

When Westergren talks about Pandora, everything seems huge, expressed in millions. Fifty million registered users growing by 85,000 per day. A library of three-quarter million songs. Ninety million songs played each month. Over 150 million session starts each week. Estimates of 70 million custom stations created.

Drilling down one realizes how minimally Pandora touches the average listener. Based on these numbers, it turns out that the average user signs on perhaps three times a week and the average session lasts for just seven songs. That’s it.

One thing that seems to really rub Westergren the wrong way is the SoundExchange royalties Pandora pays. He takes every opportunity to point out that Pandora pays millions to SoundExchange while local radio stations pay nothing.

Westergren recently bragged to Digital Music News that Pandora was about 44% of Internet radio. One might assume that he was talking about market share, but it turns out that he was actually talking about SoundExchange royalty fees. According to Westergren, Pandora pays roughly 44-45 percent of the royalties SoundExchange extracts.

As the comment was picked up by others, the message gained weight. Westergren’s comments appeared under headlines like Pandora has 44% of Internet Radio’s Audience, and Pandora is ‘about 44% of internet radio’ revenues, says founder Tim Westergren. Blame sloppy reporting, but Westergren’s misleading comment must share some of the blame.

Westergren says that Pandora paid $28 million to SoundExchange in 2009. Like other large Internet radio stations, Pandora pays 25% of gross revenue in royalties. In 2009 that should have been $12.5 million. So why does he say the company paid $28 million ($30 million in some interviews)? Maybe because $28 million sounds a lot more onerous than $12.5 million.

Someday Pandora will go public. Westergren and those who financially bankroll Pandora will become a multi-millionaires. Understandably, there is an incentive to talk up Pandora and make it seem bigger and more important than it actually is.

Fair enough.

However, since he seems to have a compulsion to distort and manipulate the facts repeatedly attacking local radio,  it is time for local radio to push back and set the record straight.

— Glenda Shrader Bos & Richard Harker of Harker Research