Pandora’s business is primarily ad-supported. Ad revenues account for 85% of total revenues. They also sell subscriptions for $36 a year that enable users to remove the ads from the Pandora app, which account for the other 15% of revenues. Until recently, the subscription package offered two main benefits- the removal of a 40 hour a month listening cap, and no ads. Pandora recently removed the listening hour cap for free users. While the subscriber base is still growing in line with total revenues, Pandora seems to think they will have to employ a primarily ad-based model for their business to work.
Pandora’s potential to scale is limited due to the nature of their royalty payments to SoundExchange, a non-profit entity that collects the payments on behalf of the record labels. Pandora must pay a fee per song streamed to each user (currently $.00102 for non-paying users, but it adds up quickly!). As P’s active user base grows so does their listening hours and content acquisition costs. In their last quarter ended July 31, revenues were up 117% year over year, but content acquisition costs were up 130%. In the last four quarters content acquisition costs have been 48%, 50%, 57%, and 50% of revenues. Pandora can’t seem to raise ad revenues faster than content costs. At some point we should see some benefits to scale on the G&A and R&D lines, but they currently only account for 15% of revenues.
It seems the only paths to substantial profitability for P are to either charge more per CPM (ad rate per one thousand impressions) or sell more ads per hour of music streamed. Getting more per CPM will probably be difficult. The high dollar CPMs are in video and display ads, but the fastest growing component of Pandora’s user base is mobile phones where there is minimal screen engagement (listening while driving etc.). Pandora will probably have to rely more heavily on audio ads going forward, which command much lower CPMs (although Pandora can offer much more targeted demographics to advertisers than terrestrial radio).
That leaves increasing the number of ads per hour. The hope for Pandora bulls might lie in a few lines tucked into their S-1:
…on average, we currently run less than one minute of audio ads per hour of listening. By comparison, the three major radio networks averaged approximately 13 minutes of ads per hour in 2009, according to SNL Kagan.
Pandora has room to dramatically increase the number of audio ads on their service, at least in comparison to terrestrial radio. Not surprisingly, Pandora’s CEO recently said they will be gradually increasing the number of audio ads in their inventory although they are leaving them at three per hour for now. However, Pandora is competing not only against terrestrial radio but also against the slew of other internet radio offerings. There is probably a limit on how many audio ads they can put on before listeners flee.
What is the optimistic scenario? Let’s assume that Pandora can double their ad inventory with the increase in audio ads. They don’t release their split between visual and audio ads, but this will be much more than a doubling of the current audio ads. And let’s assume they can hold their current average CPM rate (probably not likely given audio ads are much lower CPM than visual). So they would double ad revenues. In the last quarter that would have taken ad revenues from $58.3 million to $116.6 million. Their marketing costs in the quarter were 22% of revenues, so we will assume that 78% of that incremental $58.3 million will fall to the EBIT line. That would be $45.5 million in incremental EBIT plus the $1.5 million they actually did for $47 million in EBIT. If we annualize that we get $188 million in EBIT. That would be 12X EV/EBIT, still expensive but not totally crazy if they would still be able to grow their user base beyond the 37 million they now.
There are a myriad of things that have to go right for the above scenario to happen- among them that Pandora would need to be able to sell all of those additional ads at the current CPM rate, users would have to not flee to competitors, and Pandora would have to tightly manage the additional marketing expense. A more realistic scenario is a much less dramatic increase in ads per hour and a lower CPM rate. Let’s say P increases their ad unit sales by 50% and CPM rates fall 25%. That would only yield an incremental $17 million in EBIT per quarter. That would mean $74 million in EBIT per year for P. They would then have to triple the current user base (and EBIT) to justify their valuation. That would take the user base from 37 million to 111 million or more than 1 in every 3 Americans. (And P can’t easily expand internationally because they would have to negotiate royalty deals in each locale.)
Another massive headwind are the annual step-ups in the SoundExchange deal. P’s royalty cost per song will be going up about 8% each year until their deal with SoundExchange expires in 2015. And on top of that, they will have to renegotiate their deal with SoundExchange in 2015. It took much legal and political wrangling to get the current deal, and I would anticipate that if by some chance Pandora was very profitable in 2015 that the record labels would take much of that profit back.
It looks to me like Pandora is certainly worth far less than $14 a share, although I would love to hear other viewpoints. (It is quite interesting that the company initially expected to IPO at $7-9 while they ended up getting $16.) Timing the short is the main difficulty here. There is a decent chance P could show some earnings in future quarters off of their current breakeven base, which would give fuel to the bulls to tout P’s ”massive earnings growth”. It is hard to know at what point the story ends for a “story” stock.
Elie Rosenberg runs a value investing research website at valueslant.com. Sign up here to get his free value investing ideas and analysis by email and get his free ebook, "16 Ways to Find Undervalued Stocks."