Mel Karmazin recently compared Sirius to Pandora, bragging that Sirius makes a lot more money per subscriber than Pandora:
(Pandora) are able to generate $4.59 per year from their active users. By contrast, SiriusXM generated $141 in the year for each subscriber. I think that helps to demonstrate why we like our subscription business model.
Is this a valid way to compare Sirius to Pandora, or broadcast for that matter?
Karmazin used Pandora’s estimated active user base, but a more useful comparison is to look at Pandora’s weekly average sessions, nearly half a million in the latest reported month.
If we divide Pandora’s 2010 revenue by the service’s average sessions, it turns out that an active user is worth $467, over three times what Sirius makes from each subscriber.
So much for Karmazin’s subscription model, but how does that compare to the value of a broadcast listener?
Arbitron does not release national AQH estimates, and the current mix of PPM and diary markets makes calculating a national AQH number problematic. Nevertheless, we used Arbitron’s national RADAR estimates as a starting point.
We concluded that national full week AQH is more than four million AQH persons.
We know from the RAB that spot radio alone in 2010 totaled $14.2 billion.
This means the value of a broadcast radio listener nationally is conservatively $844, twice what a Pandora listener is worth.
Now that we have a national average, we can use this analysis to examine specific radio stations.
BIA-Kelsey recently reported 2010 station revenues for the nation’s top ten stations. If we divide the station’s revenues by their latest full week PPM AQH numbers, we can compare the stations to each other as well as the national average.
Across all top 10 stations, a listener is worth $809.44.
The graph lists the stations in order of revenue from highest on the left to the lowest on the right (click to enlarge). The height of each bar indicates how much revenue each listener generated for the station.
The horizontal line is drawn at the national $844 average.
What’s it all mean? First, Karmazin has made Sirius look good by making a misleading comparison.
Pandora makes considerably more per active user than he suggests, one-hundred times more using average sessions as the metric.
Karmazin didn’t bother to contrast Sirius with broadcast radio perhaps because he knew radio would look much better, six times better than Sirius.
Beyond these cross-platform issues, there is the question of the wide range of listener values even within the top ten stations.
Four stations fall below the estimated national average. Two are music, and two are news stations, all are New York stations.
Why are there so few music stations within the top 10 billers, and why do music stations within the top ten under-perform?
Commercial loads may have something to do with it, but these are full week numbers which should even things out a little better than Monday through Friday numbers would.
And why do a Washington, D.C. and Chicago station do so well compared to New York stations?
We’ll have much more to say in part two, but in the meantime, take a look at the graph and ponder the implications.
Then calculate what your listeners are worth.