"You better watch where you're going and remember where you've been"-I'm a simple man. Let's use Charlie Daniels advice and take a look at some of the concerns short sellers have recently raised about Pandora.
Recently a short seller named David Meyers, after a bizarre appearance on CNBC, wrote:
"If you look at Pandora's own 10-Q form, you will see it has only 3.5 million paid subscribers and a total of 76.4 million active listeners. This is only up from 76.2m since December 2013. That's a dismal 0.2% growth in six months. Ironically, Pandora boasted more listener hours this quarter. I fail to see the relevance of this. If the company doesn't have more listeners, it simply means old listeners are listening longer. I fail to see how that is at all beneficial in the long run. Maybe Pandora can squeeze in a few more commercials but that is not sustainable growth. Without more active listeners, that number will no longer increase."
Mr. Meyers, we can explain it to you-but we can't understand it for you…so try to follow along.
Year-over-year Pandora's active users grew 7.5% from 71.1 million at 6/13 to 76.4 million at 6/14. Hours grew 28.9% over the same period from 3.91 billion to 5.04 billion. This implies that hours per user per month increased from 18.5 hours per month to 21.8, up 17.8% year-over-year. Thus despite users being near an all-time high, usage is also at an all-time high-this seems very healthy to us, because it implies engagement is increasing…Engagement is essential because the more Pandora knows about its users the better it can target ads to it users. Ads with higher response rates generate higher ROI's for advertisers and attract more advertising. (The larger Pandora gets the larger the big data advantage becomes and the better their product is-for both listeners and for advertisers.)
Pandora also disclosed at the recent Needham Interconnect Conference that as of June 2014 there have been 7 million cars that activated Pandora integrated in to the dash, and that users who activate in the car listen to an incremental 6-7 hours per month (30% lift), so 10% of actives listen 30% more-driving 3% greater engagement across the base. (see Q&A from Pandora presentation at Needham Interconnect Conference, August 5th 2014) As more actives have access to Pandora in their cars, where traditionally more than 44% of radio listening takes place, engagement will gain a further boost and hours will grow. Thus without adding any active users, Pandora can increase the hours it streams and dramatically increase its revenue and earnings. To simplify, for illustrative purposes: 1) Over time Pandora will be available to its active listener base in the car, let's assume a 30% lift to hours from this (it is more likely to approximate 50% over time, but let's use current metrics to be conservative); 2) let's assume Pandora increases ad load by 50%, over time, (still way below terrestrial radio ad loads); 3) only 20% of Pandora revenue comes from local, this will increase to 70% over time-hyper local ads local ads are sold at 4x the CPM for network national, but let's assume that the shift from 20% local to 70% local results in a 2x CPM lift for 50% of P's inventory so a 50% lift across its entire inventory-these 3 steps triple Pandora's revenue, without adding active users, completely debunking Mr. Meyers assertion that "more hours" are not relevant or beneficial in the long run. Importantly, raising ad load and increasing local mix to raise the average CPM does not impact Pandora's cost per hour and would create significant gross margin leverage and earnings power. All this is hypothetical, because Pandora will grow actives and push margin expansion out further into the future, but will increase its earnings capability. So active user growth is good, but lack of active user growth is not bad, it's merely sub-optimal. Pandora's active user growth correlates with device growth, so comparing Actives in December, rather than year over year would be misleading (which is probably why Mr. Meyers chose 12/13, rather than 6/13). We are confident that a new iPhone this fall will benefit Pandora's active user count as will the holiday season.
Mr. Meyers goes on to compare Spotify's global metrics to Pandora's domestic metrics to imply Pandora's share of US internet radio market share is no longer at 77%. He goes on to point out that
"several representatives of IHeart Radio confirmed (to him) the company has 50 million active listeners. This is not a typo. Assuming ITunes Radio, GooglePlay, and Rhapsody have simply maintained their minor market share of 13%, this leaves Pandora with 35%-40% market share. This means that Pandora has lost 53% of its market share in six months"
No, it absolutely does not-all it means is that Mr. Meyers is capable of comparing apples to oranges and performing fuzzy math calculations. 1) he is mixing interactive and non-interactive services together; 2) he is giving Spotify credit for its global active user count as if its 40m actives were competing with Pandora in the US (and for US internet radio share) 3) he is assuming each service has the same engagement/active (they do not) 4) he is confusing the IHeart Radio Network (which includes non-radio sites) stats with the actual radio product that "competes" with Pandora (IHeart.com). To clarify, based on June comScore Media Metrix data, Pandora had 81.8 million unique visitors (up 27% y-o-y), while Spotify had 24.2 Million (up 19% y-o-y), and iheart.com had 17.2 Million (up 26% y-o-y). (see comScore June 2014 Media Matrix data)
Ultimately the market share that matters is revenue market share. During Q2 2014 CBS recorded radio revenue down 3% year over year, Clear channel also reported domestic radio revenue that was down year over year Pandora's revenue was up 38% year-over-year. While Pandora is building off a smaller base, it is certainly taking market share and is highly likely to continue to do so.
Finally, to illustrate Mr. Meyers really has not done his homework he writes,
"let's reiterate: Pandora lost over half its market share just in the last six months. It is limiting 'free listen' members to 40 hours a week, this limiting listener hours overall, and the company will be raising fees to potential new subscribers it is not getting. What does the company say to this?"
Fortunately, the company did not bother to respond, but for fun allow me to, line-by-line:
- "So, let's reiterate: Pandora lost over half its market share just in the last six months." As we have pointed-out above, this is factually incorrect.
- "It is limiting "free listen" members to 40 hours a week, thus limiting listener hours overall . . ." Again, this is factually inaccurate as the listening cap was removed in 2013.
- "[A]nd the company will be raising fees to potential new subscribers it is not getting. What does the company say to this?" This again is not true-the price increase occurred in q2 and Pandora added net new subscribers.
While we have chosen to single out Mr. Meyers "research" today we are grateful to other short sellers for helping to create confusion about what's really going on with Pandora and permit us to add to our position at attractive prices. We have found the following to be the most repetitive:
Shock exchange: "based upon the 3x revenue multiple apple paid for Beats, Pandora is worth $11-$12 per share, less than half the $26 price it currently trades for". At least, unlike Mr. Meyers, his math is correct, but the logic of applying revenue multiples is flawed. Beats is a headphone company with a minor interactive music service. Applying revenue multiples is reckless and comparing a transaction that involves a technology hardware company is neither comparable nor relevant. Shock Exchange may as well argue in his next post that 4 out of every 5 dentists recommend trident for their patients who chew sugarless gum, and because it's not 5 out of 5 you should short Pandora…or better yet, how about apply the WhatsApp revenue multiple to Pandora, because FB and Pandora are both pursuing local ad revenues and a merger would make sense.
Our favorite poster is Quoth the Raven, not only because we are big fans of Edgar Allen Poe, but also because, like Mr. Meyers, Shock Exchange and Rich Greenfield before them, his "research" is flawed.
"Pandora mentioned on its Q2 conference call (transcript) it now has 7M unique users who activated Pandora "native automotive integrations", up from 2.5M a year ago. However, automotive listening currently averages 6 hours/month, well below the 21.1 hours averaged by Pandora's 76.4M total active listeners."
QTR is missing several points and reaching a flawed conclusion 1) nearly all 7M unique users are already Pandora unique users, Pandora was merely illustrating the 30% lift it is getting to engagement from auto integrations-the other point QTR overlooks is the very high attachment rate Pandora is generating, the % of people who activate their native automotive integration is very high (see Needham Interconnect Conference presentation & Q&A).
"Pandora's active listener metrics have floundered, yet the stock price holds up". Perhaps this is because long term investors have built discounted cash flow analyses and know that the current price for Pandora's shares trades at a large discount to fair value, and especially when the stock trades below $25 these investors accumulate, recognizing the attractive risk/reward balance. QTR timed a couple reports with the broad-based sell-off in highly valued growth companies earlier this year and now he/she thinks his/her call on Pandora is correct, or better still that his/her reports were responsible for the movement in the stock price. In the long run, he's wrong and his posts don't matter. His reliance on iTunes Radio and Spotify to "tighten a noose" around Pandora defy evidence to the contrary. Since iTunes Radio launched, and since Spotify launched --and then relaunched and then relaunched -- Pandora has increased its active user count and increased engagement per user.
Rather than disparaging Michael Graham and other sell side analysts who attempt to do the best they can in handicapping Pandora's chances of success, QTR should build a model, like every other sell-side analyst has and he/she should publish his/her DCF. Until QTR is willing to build and publish his/her DCF and demonstrate how he/she gets to a sub $10 stock price his/her posts are not relevant.
We are very comfortable Pandora does not have an active user problem, and expect gross profit to grow meaningfully in Q3 and inflect in Q4. We accumulated Pandora's shares in 2012 below $10 and have added in 2014 below $30.
Disclosure: The author is long P. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it. The author has no business relationship with any company whose stock is mentioned in this article.