PANDORA MEDIA INC P S
February 14, 2015 - 6:37pm EST by
baileyb906
2015 2016
Price: 15.44 EPS .24 .37
Shares Out. (in M): 219 P/E 64 42
Market Cap (in $M): 3,380 P/FCF 48 31
Net Debt (in $M): -459 EBIT -69 -52
TEV ($): 2,921 TEV/EBIT NM NM
Borrow Cost: General Collateral

Sign up for free guest access to view investment idea with a 45 days delay.

  • Slowing growth
  • Competitive Threats
  • Cyclical

Description

 
Summary recommendation rationale:
 
There is clearly a lot of potential value in the brand and its giant installed user base, but the industry structure here – especially on the cost side  - is troublesome. Unlike terrestrial radio, which does not pay royalties to the music publishers for the right to play music, Pandora needs to compensate music publishers on a per play basis. The per play rate is established every 5 years by the CRB (Copyright Royalty Board), a 3 person panel which hears arguments from industry participants and sets rates. The CRB has met 3 times before, and has never lowered rates. The last time they met, they set rates so exorbitantly high that there needed to be a workaround side agreement to operate under, or else every industry player might have gone bankrupt. Expectations are high for a more favorable resolution at the latest iteration, called Web IV, the results of which will be announced this December. As the bulls will point out, there is a ton of opportunity to improve monetization here, particularly as the company invests in its local sales force in major markets, allowing it to better penetrate that higher rate segment of the business.  But one media buyer I consulted suggested that Pandora seems locked into thinking about the company as a substitute for or disruptor of terrestrial radio, when the more fruitful business strategy might be to think about the creation of products that solve problems in digital advertising. Instead of focusing on the $15 bn terrestrial radio pie that is in slow secular decline, the buyer suggested Pandora might do better focusing the $19 bn digital advertising market.
 
P is a short at $15.44 because of three major risks that you are not adequately compensated for at current levels:
 
  1. A bad ruling by the CRB at Web IV could kill the future operating leverage story and drastically lower the value of the company, whether using a DCF or a multiple of out year earnings.  While it is low probability that something catastrophic happens, there is a not insignificant tail risk here.
  2. The law of large numbers has hit them and monthly average users and hourly usage have become more difficult to consistently come by. While Q4 did show an uptick in monthly average users after several quarters of decelerating increases, gains in hourly usage (i.e., improvements in engagement) have consistently become more and more difficult to come by. Even though there are other ways to drive growth (i.e., grow monetization instead of users and hours), and even though the stock is down a lot, it is still at a “growth” valuation (37x 2015e EBITDA and 18x 2016e EBITDA) and growth investors have shown limited tolerance for these kinds of slow downs in user/usage metrics at other growth companies. When a company isn’t meaningfully profitable and the entire investment thesis rests on future operating leverage applied to a large TAM (total addressable market), any downward revisions to the TAM expectations can be very detrimental to the stock, even if the revisions are relatively minor.
  3. The long runway for monetization versus terrestrial radio is a bit of a fallacy because the rates they are getting are already a lot closer to those of terrestrial radio than they represent, especially when you adjust for the differing levels of engagement in the day parts Pandora is offering versus what terrestrial is offering. The Pandora audio ad isn’t apples to apples with a terrestrial radio ad. Given what Pandora can offer today, that might mean terrestrial is too high a benchmark. Given what Pandora could conceivably offer in the future, terrestrial might be too low a benchmark.

Given the three business risks highlighted, which will all be discussed in more detail below, I don’t have high confidence that Pandora can achieve long-term street expectations.  Expectations are for EBITDA go up 12x between 2014 and 2017 thanks to 24% topline growth over three years combined with EBITDA margins that are supposed to expand from 6.3% to 17%ish. I have reservations about both assumptions, but more reservations about the operating leverage being realized to this dramatic an extent. This long-held skepticism over the long-term final margin potential was only reinforced by the severe EBITDA margin guidance disappointment for 2015. The street was looking for margins to go from 6-7% in 2014 to 10-11% in 2015. Instead they guided to flat margins on 28% topline growth. So they can’t get any operating leverage on still dramatic topline growth this year, and after this year, content costs become more of a wild card, and topline growth could become harder to maintain, either because of increasing competition, or the law of large numbers/market share, or both. I do think revenues will grow strongly over the next few years, it’s just possible that they grow slightly below expectations or take a pause at some point due to competition before re-accelerating when connected cars really take off.  Even with the stock down dramatically off the highs, it is still trading 10.5x a 2017 consensus EBITDA number that I think has a lot of risk to it.  That is a high multiple given the lack of current profitability, the tail risk around Web IV, and the aggressive operating leverage and monetization assumptions underlying street estimates.

 

 

 

 

Brief Company Summary:

 

Pandora Media, founded in 2000, is the largest internet streaming radio company, with over 250 mm registered users and 81.5 mm monthly active users. Pandora enjoys its dominant market position as a result of a first mover advantage as well as platform ubiquity – it is available on virtually all popular consumer electronic platforms (iOS, Android, Desktop, etc.). It has also benefitted from an emphasis on an ad-supported service, as consumers have shown some meaningful price sensitivity around music subscription services. Around 80% of Pandora’s revenues come from its ad-supported free services, and the other 20% come from subscription fees on its ad-free service. Pandora offers stations in various music genres and also creates custom stations for users based on their favorite artists or songs. Users have the ability to give feedback on songs they like and don’t like, and over time this will lead to the customized stations offering them music that better suits their preferences. Pandora has made a substantial investment in the Music Genome Project, a giant database that defines songs based on a number of attributes. Cross-referencing songs that users like versus songs in the Music Genome Project database allows Pandora to help users discover music new to them that is similar to music they already know and like. Pandora has limited international presence in Australia and New Zealand, and over time would like to further its international reach, although these plans are longer-term and unlikely to be realized in the next year or two.

 

 

Bullish Investment Characteristics:

 

  1. .     Large First Mover Advantage and Dominant Market Share and Brand: 250 MM registered users is equal to 80% of the US population and one of the largest registered user bases of any service one can think of. 81.5 mm monthly active users is also a huge number…to put it in context…Netflix only has 39 mm domestic US streaming subscribers as of 4Q2014.  Despite the emergence of an ever-increasing number of new market entrants, Pandora has managed to maintain and even grow its share of streaming audio.
  2.      The Market for Streaming Audio is Growing Rapidly: Audio consumption patterns are evolving, and streaming audio is still in its infancy as an industry. The vast majority of audio consumption is still through terrestrial radio. It’s early days, and streaming audio will continue to take a greater share of the audio consumption pie, and since Pandora is the biggest piece of streaming audio, it will grow as streaming audio’s overall share grows.
  3.    Internet Radio is Under monetized Relative to Terrestrial Radio: The Terrestrial Radio Ad Market was $15 bn in 2014 and Pandora’s advertising revenues for 2014 were just $730 mm. Pandora represents approximately 80% of internet radio listening and approximately 9% of overall radio listening. If you add up internet radio and terrestrial radio ad dollars, P is probably getting <5% of the dollars for 9% of the listening. If you back out sports and talk radio, an often-quoted statistic is that they get 2% of the dollars for 9% of the listening.
  4.    Opportunity to Turn Data into Products for Advertisers and Artists: Pandora has a ton of data about their customers. They know their age, gender and zip code, among other things (like what bands they like and don’t like).  There is a ton of Big Data opportunity in their database of information, for artists and for advertisers. They have spoken a lot recently about their AMP, Artists Marketing Project, in which they will share local user data with artists and publishers, to help them, among other things, better plan their tour dates, set lists, etc. They also can use this data to help advertisers really target whom they want to reach. The real value in what they can offer, as explained by a media buyer, is not in replicating the radio “one to many” model, but instead focusing on “one to one”. Targeting is very hard to do, but if you get it right, it makes your service extremely valuable, which is one of many reasons that the simple comparison of P’s RPMs to terrestrial CPMs is an inherently flawed analysis.
  5.    The Connected Car is a Huge Long-Term Opportunity: Over 50% of audio listening happens in the car and rates for advertisements during the morning drive time are among the very highest spots in terrestrial.  While a lot of people have the ability to listen to Pandora in their car now by tethering it to their sound system with a cable, not very many people do it. Dashboard integration is crucial to adoption of Pandora in the car. Once it is in the car, the opportunity is two fold.  Listening hours/usage should go up a lot, increasing inventory for sale, leading to revenue growth. Also this inventory should be high value, and demand for it should be greater than the average Pandora inventory, driving upward pressure on rate, also supporting growth. This in my mind is the #1 slam-dunk opportunity for the company; however, I don’t think it moves the needle for them until several years out. There are approximately 250 mm registered vehicles in the US, and the annual average SAAR is 15mm or so. It suggests that it will take 16-17 years for the whole auto base to turn over, although the turnover time for primary vehicles is probably about half as long.  The design lead-time to get into the dashboard at auto OEMs is also long. While P is getting design wins, from win to show room floor, it’s at least 3 years. As evidence of how truly long term this opportunity is, XM and Sirius launched their satellite radio offering in late 2001, and three years later total industry satellite subs were just under 4 mm.  The connected car is indeed a very big opportunity for Pandora and other streaming audio players. I don’t think it really affects the income statement meaningfully until 2018 and beyond. It could probably move the needle substantially if you built out a 10 year model and drove a DCF off that, but in practicality, if P misses numbers in the short term because of the CRB decision or lack of monetization, the long-term yet to be realized opportunity of the connected car won’t offer much protection on the downside.  Also, the competitive dynamic is fluid enough that by the time the connected car is here, the competitive field could have changed. For this reason, I think that as long as P is making numbers and things are going well, they will probably “get credit” for the connected car. If they are making numbers, then their competitive position and share are in tact, and the leverage and monetization story is still in place, so the connected car offers long-term upside. If they miss revenue because they lose share to competition or can’t monetize, or they deleverage because of rising content costs, no one will care about the connected car because either it will come under question that they are the ultimate winner there or that they can make any money there. So it’s long way of saying that it is additive to the valuation in good times but not protective to the valuation in bad times.
  6. International: Company has a presence in Australia and New Zealand and say long-term they want to expand internationally but it is not a 2015 or likely a 2016 thing. The rights situation is complicated, and they are very behind Spotify, and even Rdio, so I am skeptical of this opportunity.
  7.  M&A: Given the relatively small market cap and totally dominant market share, the company could be a potential take out candidate for a company looking to get into the space. Potential buyers could be consumer electronics hardware players looking to support their platform, traditional media companies, or internet companies. Given GOOG just bought Songza and AAPL just bought Beats, it seems like MSFT is the most logical CE buyer. Most media companies have been net sellers of their music businesses, but there are companies that have been buying content aggregators/distributors, most notably YHOO, which is flush and will soon have even more cash from BABA, and once upon a time was a dominant force in streaming audio. The main impediment to M&A here though may be the potential anti-synergy derived from how many of the royalty deals have been written in the past. In prior Webcasting deals, rates have been the greater of a minimum royalty per play or a minimum % of firm revenue.  Obviously if a minimum % of firm revenue remains a fixture of these deals, then it would preclude a larger entity from buying Pandora.

 Bearish Investment Characteristics:

 

1.     Bull Case is Heavily Reliant on Realizing Major Content Leverage over the Next Few Years but the CRB Decision is a Crap Shoot: The streaming audio industry has a very unique and particular process for determining the royalties that the streaming companies pay to the publishers. Every 5 years a 3-person arbitration board called the CRB sets the rates. That process began in 2014 and will run through 2015 with a decision expected in December. A number of webcasting companies submitted rate proposals and supporting documents arguing for modest (Pandora) to dramatic reductions in what webcasters pay to SoundExchange, an entity that represents and negotiates on behalf of the three large music publishers (Sony, Universal, and Warner). SoundExchange on the other hand proposed a near doubling of the rates that they are paid. This is the fourth such Webcasting negotiating period so it is being referred to as Web IV. The rate the CRB is supposed to come up with is supposed to reflect what a willing buyer would pay a willing seller. In past Webcasting negotiations, a dearth of deals in the non-interactive streaming audio space (which is specifically defined and P falls into – the requirements include things like not being able to request a specific song or album or artist, limits on the number of times you can hear a certain artist in a given time period, limits on number of skips, etc.) led the CRB to look at interactive/on-demand deals (for Spotify-type services, where you can request an artist, album, etc.) and then calculate and apply an interactivity discount. In past Webcasting negotiations, the CRB came out with rates that were too high, and so uneconomical as to potentially drive non-interactive webcasters out of business, so side deals were later arranged, superceding the CRB decision. This time, people expect the CRB decision to stand. There are also non-interactive deals to look at this time. The Pandora deal with Merlin (a consortium of indie labels) sets an attractive royalty precedent, but the Apple/iTunes Radio deal sets a precedent that is far less favorable to Pandora and the other webcasters.  The Apple deal notably set a minimum fee of 45% of streaming revenues. Such a clause if included in the CRB decision would destroy Pandora’s earnings growth prospects because it would preclude operating leverage on the content line. There are a number of other arguments being put forth by both sides, and I could literally fill 10 pages describing them all, and the history behind all these contorted negotiations, but suffice to say this thing is a jump ball. On the one hand, there are three new people appointed to the CRB and the industry has been around longer now, and hopefully the CRB shouldn’t make any decisions so discriminatory that would be putting anyone out of business like last time. But on the other hand, there is no precedent for royalty rates ever going down, and the street bull case appears to be 100% predicated on massive leverage on content cost. If you talk to Pandora, they are sure they get some rate relief. If you talk to a record company, they are sure they get a rate increase. They are both biased. One high-ranking industry source described the likely outcome to me as completely unknowable and not analyzable. The worst case scenario – the SoundExchange proposal is adopted and rates per play double from here – is absolutely catastrophic to Pandora margins and would cause the company to go EBITDA and cash flow negative.

 

2.     Easy Growth is Already Behind the Company as User Growth Has Slowed Considerably: After experiencing rapid growth in monthly active user – 82% in 2011, 62% in 2012, and 31% in 2013 – things have notably slowed down. In Q1 2014, active users grew 8.0%. In Q2 they grew 7.5%, and in Q3 they grew only 5.2%, ending at 76.5 mm. Things did however pick up in Q4 with user growth of 7.0%. Clearly the law of large numbers is kicking in, and the company has acknowledged that while the long-term opportunity remains 100 mm monthly active users, that is a long-term goal, and making the connected car a reality is part of getting there (so the 100 mm is indeed years away). They have shifted their focus to increasing engagement despite having to actually limit total monthly listening hours per subscriber to 40 at one point in 2013 because increasing use led to increasing royalties that in the presence of subpar monetization led to a noneconomic service. Now that monetization is better the cap is lifted, the focus is on engagement and more hours per month per user. This makes sense because more hours and engagement likely means a more loyal subscriber, and it also eventually means more inventory/impressions to sell. Unfortunately, usage gains have really slowed down in 2014 relative to prior years as well, which also accounts for some of the recent stock weakness. The normal Q1 to Q2 slowdown was a little more pronounced than usual, and Q3 did not bounce back to Q1 levels as expected. Pandora management had guided to a better usage number in Q4, but then failed to deliver on that promise. And while it makes sense that at this point in the life cycle to move from adding users to increasing engagement, this doesn’t come cheap. They have initiated consumer facing marketing efforts for the first time, geared at increasing hours and also activating lapsed users. This could be a headwind to margin expansion.

 

3.     Opportunities for Rate Growth May be Overstated Because Internet and Terrestrial Comparisons are Not Apples to Apples: Pandora likes to talk about how they are 9% of industry listening but 2% of industry revenues. This makes a lot of sense conceptually but the inventory they are selling is not apples to apples comparable. Even taking sports and talk radio out of it, terrestrial radio rates tend to be highest during drive time, when the listener is captive and attentive in the car. Also, some of the highest priced terrestrial spots are the inline advertising, basically product placements (usually for local establishments) worked into the conversation by popular and trusted radio personalities. Pandora doesn’t really play in the drive time day part yet, nor does it offer any inline advertising spots.  Pandora also says that they are monetizing around low $40s RPM now. Assuming $40 per 1000 hours and 3 minutes per hour that would be a $7 CPM per 30-second spot. This is admittedly a flawed analysis because it’s not how media buyers really price internet radio – they buy demos not total audience – but using it for comparison’s sake, the average CPM for terrestrial is only $2, and a good spot CPM is considered $9ish. It’s not clear to me that Pandora is that under the terrestrial market, although I do think they do have a far greater opportunity than just supplanting the terrestrial market if they can really harness the power of their data.

 

4.     Growing Competition: The number of competitors participating in the streaming audio space has increased greatly. This is clearly evident just by looking at the increased number of companies participating in Web IV. The biggest threat competitively to Pandora would be the growing popularity of Spotify, particularly in the younger demo. While Spotify is still much smaller than Pandora with an estimated 60 mm subscribers (15 mm of them paid subs), it is growing users much faster, is getting linked to a lot in mainstream media because of its playlist features, and is way ahead of Pandora internationally. It has also been very aggressive on technologically innovating. While only directionally indicative, Google Trends shows search activity for Pandora falling 20% from Dec. 2012 to Dec. 2013, then another 24% from Dec. 2013 to Dec. 2014. Conversely, Google Trends shows search activity for Spotify increasing 25% from Dec. 2012 to Dec. 2013, then increasing 67% from Dec. 2013 to Dec. 2014. On the other end of the spectrum, there are a bunch of well-funded giants who have yet to make much impact but clearly have deep pockets. Much was made of the Apple iTunes Radio launch, although it hasn’t eaten into Pandora’s market share at all. Apple did however buy Beats a year ago and has yet to do anything with its small music on demand service. While Pandora has made a couple of industry hires to try to become a better promotional partner to the music publishers, it was widely thought at the time of the Apple/Beats deal that one of the motivations behind the deal might have been to bring Jimmy Iovine (and to a lesser extent Dr. Dre) in house at Apple to build a bridge to the publishers and really transform the music industry. iTunes Radio is obviously not platform agnostic (iOS only) and it’s unclear if any future Beats innovations would continue to run on Android as well as iOS (seems unlikely), which would help protect Pandora. Google, already a big force in music through YouTube, recently bought Songza. Amazon recently launched streaming music free with Amazon Prime.

 

5.     Losing Their Cool/Problems with the Younger Demo: Anecdotally, all the younger people seem to be on Spotify. There is some feeling that Pandora is your parents’ internet radio service, although the conventional wisdom of a service’s imminent demise because teens are abandoning it seems perennially overdone (see FB in the $20s). That said, the media buyer I spoke to said they specifically use Pandora to buy the 18-24 demo, so if that demo is leaving for Spotify, that will be a problem for Pandora monetization.,

 

6.     Not Enough Songs?: Pandora Catalog is 900,000 songs versus Spotify catalog is 20 million songs.

 

7.     Potential for One-Time Negative Event Concerning pre-1972 Royalty State Lawsuits: Due to pre-1972 recordings falling under state laws and not federal statutes, they have not fallen under the Webcasting agreements for minimum royalties and they have been played free for some time. There are a number of cases challenging this moving through the state courts, and so far the music publishers have been winning. Eventually Pandora will run out of appeals. If they lose, they may have to start paying royalties on pre-1972 recordings, which constitute 8-10% of plays, and have been played for free up to now. Assuming the current rate structure stands, this could be a several hundred bp margin headwind, not insignificant to a company with a 6-7% EBITDA margin. Additionally, they could be liable for 3 years of damages, which I estimate could run somewhere between $20 mm on the low end to  $140 mm on the high end. On the low end, the damages are immaterial. On the high end, it’s a third of their cash. I think the interesting thing is no one is talking about this issue. I am not certain when these cases wrap up.

 

8.     International Opportunity May Be Limited: Spotify has a big lead already, and the royalty rates have been hard enough to work out just in the US. If Pandora management thought it would be easily tackled, why would they be pushing off even framing the opportunity to 2016?

 

9.     General Macroeconomic Risks: Advertising is highly cyclical, and even when the economy is doing OK, can be lumpy and surprise people (as it did over the summer in both the television and terrestrial radio markets). While Pandora is a story of secular growth, if the overall advertising pie shrinks, it will be harder for them to grow.  This has never been considered a risk factor for them before because of their secular growth, although this may have changed some with the Q4 report.

 

 

 

Miscellaneous Observations:                  

 

1.     Despite terrible trailing twelve month performance and this very bimodal risk in the December Web IV decision, the street remained near uniformly bullish until the Q4 blow up. There were over 20 buys and 5 holds going into the miss and bad guidance. Three firms did downgrade to neutral after the quarter, and almost everyone cut their price target, but most analysts stuck with the buy, and Pandora still has 19 buy ratings on it. It’s unusual to see a stock down over 50% in a year and the sellside stick by it like this.

 

2.     Short interest which was once insane here (>70% of float I think a couple of years ago) is now down to 14%.

 

3.     Stock based compensation levels were already high here and are increasing dramatically in the face of poor performance. It is egregious.

 

 

 

Price Target and rationale:

 

Price Target of $11 with a bias, acknowledging there is a chance it is worth a lot less

 

 

 

25% chance the street is right: They get flat or slightly down rates at Web IV, monetization continues to improve, and increasing competition doesn’t eat into their market share. P makes a high teens EBITDA margin and $1.20 eps in 2017 (untaxed), a generous 20x EBITDA multiple, discounted back 2 years at 10%, yields $20

 

 

 

65% chance the street is being a little aggressive: The CRB comes back with modest increases in the royalty rate, which curtails their ability to get all the leverage the street is looking for. The benefit of RPM (revenue per thousand listening hours) growth that exceeds LPM (licensing fees per thousand plays) growth is partially offset by having to start paying royalties on pre-1972 music, which constitute about 10% of plays, and which they currently don’t pay for. Leverage on the content line is still there but more muted and EBITDA margins go from 6.3% in 2014 to 11.5% in 2017. Sales growth compounds around the recently lowered consensus numbers, as the street has woken up a bit to the law of large numbers, and I do acknowledge the very real opportunities for better monetization with the increased investment in the local sales force. P makes an untaxed 55c in 2017, and 10x 2017 EBITDA is around $9.

 

 

 

10% chance they are screwed by the CRB in December: The CRB comes back with an adverse decision – accepting the SoundExchange proposal or something close to it. P also loses the pre-1972 cases and experiences negative operating leverage on the content line. They need to throttle usage again to stay profitable. Growth rates are lowered as a result. Marketing is cut back to 2013-2014 levels through severe cost cutting to minimize cash burn. Company will be approaching EPS and EBITDA breakeven in 2017, having burned through most of its cash balance. This scenario may be too generous because RPM growth may not come through as I have modeled it in this scenario, as marketing investment will grind to halt, and the company will irritate its users by bringing back usage caps at the same time it is under threat from new competitors. The company is really near worthless in this scenario, and will probably have been forced to do a dilutive secondary to shore up its balance sheet. I will give it the benefit of the doubt and say it is worth $2 in this scenario. I do think this is an unlikely scenario, but it is catastrophic if it happens so it can’t be ignored.

 

 

 

(25% X $20) + (65% X $9) + (10% X $2) = $11.05

 

 

Risks:

They already lowered the year so much, it's kind of a sandbag, so that sets them up for possible squeezes.

They can sign favorable royalty deals with small, meaningless content owners which lulls people into complacency about their bargaining position with the Big 3 publishers.

Dream scenario happens where CRB cuts rate for them and margins explode.

 

Key Upcoming Events:

 

·      2/17/15 Written rebuttal testimony die for Web IV proceedings

 

·      3/5/15 Company will hold an Investor Day (announced on 2/5/15 with earnings)

 

·      3/20/15 Conclusion second discovery period of Web IV

 

·      4/7/15 Due date for amended written rebuttals

 

·      4/27/15 Hearings to cover rebuttals (should last one month)

 

·      Late April 2015 1Q Earnings Release

 

·      6/3/15 Closing arguments made to CRB

 

·      December 2015 Web IV Decision

 

 

I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise do not hold a material investment in the issuer's securities.

Catalyst

  1. Continued disappointments in terms of operating margin leverage as they report future quarters, although I acknowledge that they have probably lowered expectations for the year enough already so current 2015 estimates are probably attainable
  2. Continued noise around the impending CRB decision culminating in the actual decision in December 2015
  3. Loss of the pre-1972 royalty lawsuits leading to payment of damages and the need to pay royalties on this content going forward (timing uncertain)
  4. 2016 and 2017 numbers are likely too high (so this is a missing numbers story, although not in the near term)
  5. Competitive announcements - most likely further subscriber momentum and product innovation from Spotify, plus Apple could announce a strategy for Beats at any time
  6. More articles and bad PR from popular artists complaining they are underpaid for internet radio (see royalty beefs from Taylor Swift, Pharrell, etc.), there was even some griping recently on the Grammy telecast
  7. Major artists start pulling content out of the master agreements and demanding to cut higher priced side deals (see Led Zeppelin and Spotify), or withhold content altogether from streaming
    sort by    

    Description

     
    Summary recommendation rationale:
     
    There is clearly a lot of potential value in the brand and its giant installed user base, but the industry structure here – especially on the cost side  - is troublesome. Unlike terrestrial radio, which does not pay royalties to the music publishers for the right to play music, Pandora needs to compensate music publishers on a per play basis. The per play rate is established every 5 years by the CRB (Copyright Royalty Board), a 3 person panel which hears arguments from industry participants and sets rates. The CRB has met 3 times before, and has never lowered rates. The last time they met, they set rates so exorbitantly high that there needed to be a workaround side agreement to operate under, or else every industry player might have gone bankrupt. Expectations are high for a more favorable resolution at the latest iteration, called Web IV, the results of which will be announced this December. As the bulls will point out, there is a ton of opportunity to improve monetization here, particularly as the company invests in its local sales force in major markets, allowing it to better penetrate that higher rate segment of the business.  But one media buyer I consulted suggested that Pandora seems locked into thinking about the company as a substitute for or disruptor of terrestrial radio, when the more fruitful business strategy might be to think about the creation of products that solve problems in digital advertising. Instead of focusing on the $15 bn terrestrial radio pie that is in slow secular decline, the buyer suggested Pandora might do better focusing the $19 bn digital advertising market.
     
    P is a short at $15.44 because of three major risks that you are not adequately compensated for at current levels:
     
    1. A bad ruling by the CRB at Web IV could kill the future operating leverage story and drastically lower the value of the company, whether using a DCF or a multiple of out year earnings.  While it is low probability that something catastrophic happens, there is a not insignificant tail risk here.
    2. The law of large numbers has hit them and monthly average users and hourly usage have become more difficult to consistently come by. While Q4 did show an uptick in monthly average users after several quarters of decelerating increases, gains in hourly usage (i.e., improvements in engagement) have consistently become more and more difficult to come by. Even though there are other ways to drive growth (i.e., grow monetization instead of users and hours), and even though the stock is down a lot, it is still at a “growth” valuation (37x 2015e EBITDA and 18x 2016e EBITDA) and growth investors have shown limited tolerance for these kinds of slow downs in user/usage metrics at other growth companies. When a company isn’t meaningfully profitable and the entire investment thesis rests on future operating leverage applied to a large TAM (total addressable market), any downward revisions to the TAM expectations can be very detrimental to the stock, even if the revisions are relatively minor.
    3. The long runway for monetization versus terrestrial radio is a bit of a fallacy because the rates they are getting are already a lot closer to those of terrestrial radio than they represent, especially when you adjust for the differing levels of engagement in the day parts Pandora is offering versus what terrestrial is offering. The Pandora audio ad isn’t apples to apples with a terrestrial radio ad. Given what Pandora can offer today, that might mean terrestrial is too high a benchmark. Given what Pandora could conceivably offer in the future, terrestrial might be too low a benchmark.

    Given the three business risks highlighted, which will all be discussed in more detail below, I don’t have high confidence that Pandora can achieve long-term street expectations.  Expectations are for EBITDA go up 12x between 2014 and 2017 thanks to 24% topline growth over three years combined with EBITDA margins that are supposed to expand from 6.3% to 17%ish. I have reservations about both assumptions, but more reservations about the operating leverage being realized to this dramatic an extent. This long-held skepticism over the long-term final margin potential was only reinforced by the severe EBITDA margin guidance disappointment for 2015. The street was looking for margins to go from 6-7% in 2014 to 10-11% in 2015. Instead they guided to flat margins on 28% topline growth. So they can’t get any operating leverage on still dramatic topline growth this year, and after this year, content costs become more of a wild card, and topline growth could become harder to maintain, either because of increasing competition, or the law of large numbers/market share, or both. I do think revenues will grow strongly over the next few years, it’s just possible that they grow slightly below expectations or take a pause at some point due to competition before re-accelerating when connected cars really take off.  Even with the stock down dramatically off the highs, it is still trading 10.5x a 2017 consensus EBITDA number that I think has a lot of risk to it.  That is a high multiple given the lack of current profitability, the tail risk around Web IV, and the aggressive operating leverage and monetization assumptions underlying street estimates.

     

     

     

     

    Brief Company Summary:

     

    Pandora Media, founded in 2000, is the largest internet streaming radio company, with over 250 mm registered users and 81.5 mm monthly active users. Pandora enjoys its dominant market position as a result of a first mover advantage as well as platform ubiquity – it is available on virtually all popular consumer electronic platforms (iOS, Android, Desktop, etc.). It has also benefitted from an emphasis on an ad-supported service, as consumers have shown some meaningful price sensitivity around music subscription services. Around 80% of Pandora’s revenues come from its ad-supported free services, and the other 20% come from subscription fees on its ad-free service. Pandora offers stations in various music genres and also creates custom stations for users based on their favorite artists or songs. Users have the ability to give feedback on songs they like and don’t like, and over time this will lead to the customized stations offering them music that better suits their preferences. Pandora has made a substantial investment in the Music Genome Project, a giant database that defines songs based on a number of attributes. Cross-referencing songs that users like versus songs in the Music Genome Project database allows Pandora to help users discover music new to them that is similar to music they already know and like. Pandora has limited international presence in Australia and New Zealand, and over time would like to further its international reach, although these plans are longer-term and unlikely to be realized in the next year or two.

     

     

    Bullish Investment Characteristics:

     

    1. .     Large First Mover Advantage and Dominant Market Share and Brand: 250 MM registered users is equal to 80% of the US population and one of the largest registered user bases of any service one can think of. 81.5 mm monthly active users is also a huge number…to put it in context…Netflix only has 39 mm domestic US streaming subscribers as of 4Q2014.  Despite the emergence of an ever-increasing number of new market entrants, Pandora has managed to maintain and even grow its share of streaming audio.
    2.      The Market for Streaming Audio is Growing Rapidly: Audio consumption patterns are evolving, and streaming audio is still in its infancy as an industry. The vast majority of audio consumption is still through terrestrial radio. It’s early days, and streaming audio will continue to take a greater share of the audio consumption pie, and since Pandora is the biggest piece of streaming audio, it will grow as streaming audio’s overall share grows.
    3.    Internet Radio is Under monetized Relative to Terrestrial Radio: The Terrestrial Radio Ad Market was $15 bn in 2014 and Pandora’s advertising revenues for 2014 were just $730 mm. Pandora represents approximately 80% of internet radio listening and approximately 9% of overall radio listening. If you add up internet radio and terrestrial radio ad dollars, P is probably getting <5% of the dollars for 9% of the listening. If you back out sports and talk radio, an often-quoted statistic is that they get 2% of the dollars for 9% of the listening.
    4.    Opportunity to Turn Data into Products for Advertisers and Artists: Pandora has a ton of data about their customers. They know their age, gender and zip code, among other things (like what bands they like and don’t like).  There is a ton of Big Data opportunity in their database of information, for artists and for advertisers. They have spoken a lot recently about their AMP, Artists Marketing Project, in which they will share local user data with artists and publishers, to help them, among other things, better plan their tour dates, set lists, etc. They also can use this data to help advertisers really target whom they want to reach. The real value in what they can offer, as explained by a media buyer, is not in replicating the radio “one to many” model, but instead focusing on “one to one”. Targeting is very hard to do, but if you get it right, it makes your service extremely valuable, which is one of many reasons that the simple comparison of P’s RPMs to terrestrial CPMs is an inherently flawed analysis.
    5.    The Connected Car is a Huge Long-Term Opportunity: Over 50% of audio listening happens in the car and rates for advertisements during the morning drive time are among the very highest spots in terrestrial.  While a lot of people have the ability to listen to Pandora in their car now by tethering it to their sound system with a cable, not very many people do it. Dashboard integration is crucial to adoption of Pandora in the car. Once it is in the car, the opportunity is two fold.  Listening hours/usage should go up a lot, increasing inventory for sale, leading to revenue growth. Also this inventory should be high value, and demand for it should be greater than the average Pandora inventory, driving upward pressure on rate, also supporting growth. This in my mind is the #1 slam-dunk opportunity for the company; however, I don’t think it moves the needle for them until several years out. There are approximately 250 mm registered vehicles in the US, and the annual average SAAR is 15mm or so. It suggests that it will take 16-17 years for the whole auto base to turn over, although the turnover time for primary vehicles is probably about half as long.  The design lead-time to get into the dashboard at auto OEMs is also long. While P is getting design wins, from win to show room floor, it’s at least 3 years. As evidence of how truly long term this opportunity is, XM and Sirius launched their satellite radio offering in late 2001, and three years later total industry satellite subs were just under 4 mm.  The connected car is indeed a very big opportunity for Pandora and other streaming audio players. I don’t think it really affects the income statement meaningfully until 2018 and beyond. It could probably move the needle substantially if you built out a 10 year model and drove a DCF off that, but in practicality, if P misses numbers in the short term because of the CRB decision or lack of monetization, the long-term yet to be realized opportunity of the connected car won’t offer much protection on the downside.  Also, the competitive dynamic is fluid enough that by the time the connected car is here, the competitive field could have changed. For this reason, I think that as long as P is making numbers and things are going well, they will probably “get credit” for the connected car. If they are making numbers, then their competitive position and share are in tact, and the leverage and monetization story is still in place, so the connected car offers long-term upside. If they miss revenue because they lose share to competition or can’t monetize, or they deleverage because of rising content costs, no one will care about the connected car because either it will come under question that they are the ultimate winner there or that they can make any money there. So it’s long way of saying that it is additive to the valuation in good times but not protective to the valuation in bad times.
    6. International: Company has a presence in Australia and New Zealand and say long-term they want to expand internationally but it is not a 2015 or likely a 2016 thing. The rights situation is complicated, and they are very behind Spotify, and even Rdio, so I am skeptical of this opportunity.
    7.  M&A: Given the relatively small market cap and totally dominant market share, the company could be a potential take out candidate for a company looking to get into the space. Potential buyers could be consumer electronics hardware players looking to support their platform, traditional media companies, or internet companies. Given GOOG just bought Songza and AAPL just bought Beats, it seems like MSFT is the most logical CE buyer. Most media companies have been net sellers of their music businesses, but there are companies that have been buying content aggregators/distributors, most notably YHOO, which is flush and will soon have even more cash from BABA, and once upon a time was a dominant force in streaming audio. The main impediment to M&A here though may be the potential anti-synergy derived from how many of the royalty deals have been written in the past. In prior Webcasting deals, rates have been the greater of a minimum royalty per play or a minimum % of firm revenue.  Obviously if a minimum % of firm revenue remains a fixture of these deals, then it would preclude a larger entity from buying Pandora.

     Bearish Investment Characteristics:

     

    1.     Bull Case is Heavily Reliant on Realizing Major Content Leverage over the Next Few Years but the CRB Decision is a Crap Shoot: The streaming audio industry has a very unique and particular process for determining the royalties that the streaming companies pay to the publishers. Every 5 years a 3-person arbitration board called the CRB sets the rates. That process began in 2014 and will run through 2015 with a decision expected in December. A number of webcasting companies submitted rate proposals and supporting documents arguing for modest (Pandora) to dramatic reductions in what webcasters pay to SoundExchange, an entity that represents and negotiates on behalf of the three large music publishers (Sony, Universal, and Warner). SoundExchange on the other hand proposed a near doubling of the rates that they are paid. This is the fourth such Webcasting negotiating period so it is being referred to as Web IV. The rate the CRB is supposed to come up with is supposed to reflect what a willing buyer would pay a willing seller. In past Webcasting negotiations, a dearth of deals in the non-interactive streaming audio space (which is specifically defined and P falls into – the requirements include things like not being able to request a specific song or album or artist, limits on the number of times you can hear a certain artist in a given time period, limits on number of skips, etc.) led the CRB to look at interactive/on-demand deals (for Spotify-type services, where you can request an artist, album, etc.) and then calculate and apply an interactivity discount. In past Webcasting negotiations, the CRB came out with rates that were too high, and so uneconomical as to potentially drive non-interactive webcasters out of business, so side deals were later arranged, superceding the CRB decision. This time, people expect the CRB decision to stand. There are also non-interactive deals to look at this time. The Pandora deal with Merlin (a consortium of indie labels) sets an attractive royalty precedent, but the Apple/iTunes Radio deal sets a precedent that is far less favorable to Pandora and the other webcasters.  The Apple deal notably set a minimum fee of 45% of streaming revenues. Such a clause if included in the CRB decision would destroy Pandora’s earnings growth prospects because it would preclude operating leverage on the content line. There are a number of other arguments being put forth by both sides, and I could literally fill 10 pages describing them all, and the history behind all these contorted negotiations, but suffice to say this thing is a jump ball. On the one hand, there are three new people appointed to the CRB and the industry has been around longer now, and hopefully the CRB shouldn’t make any decisions so discriminatory that would be putting anyone out of business like last time. But on the other hand, there is no precedent for royalty rates ever going down, and the street bull case appears to be 100% predicated on massive leverage on content cost. If you talk to Pandora, they are sure they get some rate relief. If you talk to a record company, they are sure they get a rate increase. They are both biased. One high-ranking industry source described the likely outcome to me as completely unknowable and not analyzable. The worst case scenario – the SoundExchange proposal is adopted and rates per play double from here – is absolutely catastrophic to Pandora margins and would cause the company to go EBITDA and cash flow negative.

     

    2.     Easy Growth is Already Behind the Company as User Growth Has Slowed Considerably: After experiencing rapid growth in monthly active user – 82% in 2011, 62% in 2012, and 31% in 2013 – things have notably slowed down. In Q1 2014, active users grew 8.0%. In Q2 they grew 7.5%, and in Q3 they grew only 5.2%, ending at 76.5 mm. Things did however pick up in Q4 with user growth of 7.0%. Clearly the law of large numbers is kicking in, and the company has acknowledged that while the long-term opportunity remains 100 mm monthly active users, that is a long-term goal, and making the connected car a reality is part of getting there (so the 100 mm is indeed years away). They have shifted their focus to increasing engagement despite having to actually limit total monthly listening hours per subscriber to 40 at one point in 2013 because increasing use led to increasing royalties that in the presence of subpar monetization led to a noneconomic service. Now that monetization is better the cap is lifted, the focus is on engagement and more hours per month per user. This makes sense because more hours and engagement likely means a more loyal subscriber, and it also eventually means more inventory/impressions to sell. Unfortunately, usage gains have really slowed down in 2014 relative to prior years as well, which also accounts for some of the recent stock weakness. The normal Q1 to Q2 slowdown was a little more pronounced than usual, and Q3 did not bounce back to Q1 levels as expected. Pandora management had guided to a better usage number in Q4, but then failed to deliver on that promise. And while it makes sense that at this point in the life cycle to move from adding users to increasing engagement, this doesn’t come cheap. They have initiated consumer facing marketing efforts for the first time, geared at increasing hours and also activating lapsed users. This could be a headwind to margin expansion.

     

    3.     Opportunities for Rate Growth May be Overstated Because Internet and Terrestrial Comparisons are Not Apples to Apples: Pandora likes to talk about how they are 9% of industry listening but 2% of industry revenues. This makes a lot of sense conceptually but the inventory they are selling is not apples to apples comparable. Even taking sports and talk radio out of it, terrestrial radio rates tend to be highest during drive time, when the listener is captive and attentive in the car. Also, some of the highest priced terrestrial spots are the inline advertising, basically product placements (usually for local establishments) worked into the conversation by popular and trusted radio personalities. Pandora doesn’t really play in the drive time day part yet, nor does it offer any inline advertising spots.  Pandora also says that they are monetizing around low $40s RPM now. Assuming $40 per 1000 hours and 3 minutes per hour that would be a $7 CPM per 30-second spot. This is admittedly a flawed analysis because it’s not how media buyers really price internet radio – they buy demos not total audience – but using it for comparison’s sake, the average CPM for terrestrial is only $2, and a good spot CPM is considered $9ish. It’s not clear to me that Pandora is that under the terrestrial market, although I do think they do have a far greater opportunity than just supplanting the terrestrial market if they can really harness the power of their data.

     

    4.     Growing Competition: The number of competitors participating in the streaming audio space has increased greatly. This is clearly evident just by looking at the increased number of companies participating in Web IV. The biggest threat competitively to Pandora would be the growing popularity of Spotify, particularly in the younger demo. While Spotify is still much smaller than Pandora with an estimated 60 mm subscribers (15 mm of them paid subs), it is growing users much faster, is getting linked to a lot in mainstream media because of its playlist features, and is way ahead of Pandora internationally. It has also been very aggressive on technologically innovating. While only directionally indicative, Google Trends shows search activity for Pandora falling 20% from Dec. 2012 to Dec. 2013, then another 24% from Dec. 2013 to Dec. 2014. Conversely, Google Trends shows search activity for Spotify increasing 25% from Dec. 2012 to Dec. 2013, then increasing 67% from Dec. 2013 to Dec. 2014. On the other end of the spectrum, there are a bunch of well-funded giants who have yet to make much impact but clearly have deep pockets. Much was made of the Apple iTunes Radio launch, although it hasn’t eaten into Pandora’s market share at all. Apple did however buy Beats a year ago and has yet to do anything with its small music on demand service. While Pandora has made a couple of industry hires to try to become a better promotional partner to the music publishers, it was widely thought at the time of the Apple/Beats deal that one of the motivations behind the deal might have been to bring Jimmy Iovine (and to a lesser extent Dr. Dre) in house at Apple to build a bridge to the publishers and really transform the music industry. iTunes Radio is obviously not platform agnostic (iOS only) and it’s unclear if any future Beats innovations would continue to run on Android as well as iOS (seems unlikely), which would help protect Pandora. Google, already a big force in music through YouTube, recently bought Songza. Amazon recently launched streaming music free with Amazon Prime.

     

    5.     Losing Their Cool/Problems with the Younger Demo: Anecdotally, all the younger people seem to be on Spotify. There is some feeling that Pandora is your parents’ internet radio service, although the conventional wisdom of a service’s imminent demise because teens are abandoning it seems perennially overdone (see FB in the $20s). That said, the media buyer I spoke to said they specifically use Pandora to buy the 18-24 demo, so if that demo is leaving for Spotify, that will be a problem for Pandora monetization.,

     

    6.     Not Enough Songs?: Pandora Catalog is 900,000 songs versus Spotify catalog is 20 million songs.

     

    7.     Potential for One-Time Negative Event Concerning pre-1972 Royalty State Lawsuits: Due to pre-1972 recordings falling under state laws and not federal statutes, they have not fallen under the Webcasting agreements for minimum royalties and they have been played free for some time. There are a number of cases challenging this moving through the state courts, and so far the music publishers have been winning. Eventually Pandora will run out of appeals. If they lose, they may have to start paying royalties on pre-1972 recordings, which constitute 8-10% of plays, and have been played for free up to now. Assuming the current rate structure stands, this could be a several hundred bp margin headwind, not insignificant to a company with a 6-7% EBITDA margin. Additionally, they could be liable for 3 years of damages, which I estimate could run somewhere between $20 mm on the low end to  $140 mm on the high end. On the low end, the damages are immaterial. On the high end, it’s a third of their cash. I think the interesting thing is no one is talking about this issue. I am not certain when these cases wrap up.

     

    8.     International Opportunity May Be Limited: Spotify has a big lead already, and the royalty rates have been hard enough to work out just in the US. If Pandora management thought it would be easily tackled, why would they be pushing off even framing the opportunity to 2016?

     

    9.     General Macroeconomic Risks: Advertising is highly cyclical, and even when the economy is doing OK, can be lumpy and surprise people (as it did over the summer in both the television and terrestrial radio markets). While Pandora is a story of secular growth, if the overall advertising pie shrinks, it will be harder for them to grow.  This has never been considered a risk factor for them before because of their secular growth, although this may have changed some with the Q4 report.

     

     

     

    Miscellaneous Observations:                  

     

    1.     Despite terrible trailing twelve month performance and this very bimodal risk in the December Web IV decision, the street remained near uniformly bullish until the Q4 blow up. There were over 20 buys and 5 holds going into the miss and bad guidance. Three firms did downgrade to neutral after the quarter, and almost everyone cut their price target, but most analysts stuck with the buy, and Pandora still has 19 buy ratings on it. It’s unusual to see a stock down over 50% in a year and the sellside stick by it like this.

     

    2.     Short interest which was once insane here (>70% of float I think a couple of years ago) is now down to 14%.

     

    3.     Stock based compensation levels were already high here and are increasing dramatically in the face of poor performance. It is egregious.

     

     

     

    Price Target and rationale:

     

    Price Target of $11 with a bias, acknowledging there is a chance it is worth a lot less

     

     

     

    25% chance the street is right: They get flat or slightly down rates at Web IV, monetization continues to improve, and increasing competition doesn’t eat into their market share. P makes a high teens EBITDA margin and $1.20 eps in 2017 (untaxed), a generous 20x EBITDA multiple, discounted back 2 years at 10%, yields $20

     

     

     

    65% chance the street is being a little aggressive: The CRB comes back with modest increases in the royalty rate, which curtails their ability to get all the leverage the street is looking for. The benefit of RPM (revenue per thousand listening hours) growth that exceeds LPM (licensing fees per thousand plays) growth is partially offset by having to start paying royalties on pre-1972 music, which constitute about 10% of plays, and which they currently don’t pay for. Leverage on the content line is still there but more muted and EBITDA margins go from 6.3% in 2014 to 11.5% in 2017. Sales growth compounds around the recently lowered consensus numbers, as the street has woken up a bit to the law of large numbers, and I do acknowledge the very real opportunities for better monetization with the increased investment in the local sales force. P makes an untaxed 55c in 2017, and 10x 2017 EBITDA is around $9.

     

     

     

    10% chance they are screwed by the CRB in December: The CRB comes back with an adverse decision – accepting the SoundExchange proposal or something close to it. P also loses the pre-1972 cases and experiences negative operating leverage on the content line. They need to throttle usage again to stay profitable. Growth rates are lowered as a result. Marketing is cut back to 2013-2014 levels through severe cost cutting to minimize cash burn. Company will be approaching EPS and EBITDA breakeven in 2017, having burned through most of its cash balance. This scenario may be too generous because RPM growth may not come through as I have modeled it in this scenario, as marketing investment will grind to halt, and the company will irritate its users by bringing back usage caps at the same time it is under threat from new competitors. The company is really near worthless in this scenario, and will probably have been forced to do a dilutive secondary to shore up its balance sheet. I will give it the benefit of the doubt and say it is worth $2 in this scenario. I do think this is an unlikely scenario, but it is catastrophic if it happens so it can’t be ignored.

     

     

     

    (25% X $20) + (65% X $9) + (10% X $2) = $11.05

     

     

    Risks:

    They already lowered the year so much, it's kind of a sandbag, so that sets them up for possible squeezes.

    They can sign favorable royalty deals with small, meaningless content owners which lulls people into complacency about their bargaining position with the Big 3 publishers.

    Dream scenario happens where CRB cuts rate for them and margins explode.

     

    Key Upcoming Events:

     

    ·      2/17/15 Written rebuttal testimony die for Web IV proceedings

     

    ·      3/5/15 Company will hold an Investor Day (announced on 2/5/15 with earnings)

     

    ·      3/20/15 Conclusion second discovery period of Web IV

     

    ·      4/7/15 Due date for amended written rebuttals

     

    ·      4/27/15 Hearings to cover rebuttals (should last one month)

     

    ·      Late April 2015 1Q Earnings Release

     

    ·      6/3/15 Closing arguments made to CRB

     

    ·      December 2015 Web IV Decision

     

     

    I do not hold a position with the issuer such as employment, directorship, or consultancy.
    I and/or others I advise do not hold a material investment in the issuer's securities.

    Catalyst

    1. Continued disappointments in terms of operating margin leverage as they report future quarters, although I acknowledge that they have probably lowered expectations for the year enough already so current 2015 estimates are probably attainable
    2. Continued noise around the impending CRB decision culminating in the actual decision in December 2015
    3. Loss of the pre-1972 royalty lawsuits leading to payment of damages and the need to pay royalties on this content going forward (timing uncertain)
    4. 2016 and 2017 numbers are likely too high (so this is a missing numbers story, although not in the near term)
    5. Competitive announcements - most likely further subscriber momentum and product innovation from Spotify, plus Apple could announce a strategy for Beats at any time
    6. More articles and bad PR from popular artists complaining they are underpaid for internet radio (see royalty beefs from Taylor Swift, Pharrell, etc.), there was even some griping recently on the Grammy telecast
    7. Major artists start pulling content out of the master agreements and demanding to cut higher priced side deals (see Led Zeppelin and Spotify), or withhold content altogether from streaming
      Back to top