Weekend Edition: The Best Advertisement For Satellite Radio

Your trip was a 60 min UNadvertisement for terrestrial radio. There are so many potential ways that satellite radio’s delivery mechanism can be replaced that you can only value satellite radio companies based on their proprietary content.

Here are a few illustrative technologies;

*. 802.11a or whatever comes after it has the capability to send high bandwidth digital signals over a wide area. For the price of XM’s annual depreciation you could probably outfit the top 10 metro areas and offer more channels.
*. You mentioned the ipod. You could already take a pod cast with you. How about your car radio stores 2 hours of 30 channels from your home wifi network before you leave

I may be (I almost certainly am) off on the exact technology but over the next 3-5 years there will be technology to deliver more channels much more cheaply than XM or Sirius.
Terrestrial as it stands today is dead but satellite delivery will probably never show real return on capital.

Of course XM, Sirius or any of the terrestrial companies may set up these new digital terrestrial networks but I can’t see XM or Sirius doing it because it will damage their existing business.

I don’t disagree with any of the enthusiasm for the satellite radio product, or for the growth of the industry. As Jeff mentioned in his post though, that doesn’t necessarily translate to enthusiasm for the stocks. SIRI currently trades at a premium to XM, though XM has 3x as many subscribers, and its questionable if XM is even fairly valued. Some quick math will enlighten (I’ll use SIRI as the example, since it is the more overvalued of the two):

– SIRI chareges 12.99/mo/sub for service, but in actuality gets closer to 10.50, due to deals on multi year contracts. Using the 12.99, you get annual revs per sub of $155.9, which at a 45% contribution margin means each sub adds $70/year in operating profit. Capitalize that at 6,8, and 10% and you get per sub values of 1,166, 875 and 700. At a current enterprise value of $6.4B, this implies subs of 5.49M, 7.31M and 9.14M respectively. SIRI currently has 1.2M subs (probably 1.5M at the end of Q1), but the market should be able to value out 6-12 months, so looking at year end 05 or mid year 06 numbers is more applicable. at YE 05, they will have 2.6M subs, and at mid year 06, they will have 3.9M subs. However you do the math, it doesn’t add up.

– Perhaps a per sub valuation is not the correct way to look at the company (even though thats the way DBS, cable, and other sub businesses are) and a DCF is the better approach. According to one analyst, using a 15% cost of equity, a 12.5% FCF perpetual growth rate and a 38.5x FCF multiple, the resulting share price is $4. I would argue those are some fairly aggressive assumptions and yet still yield a price 20% below today’s price.

Long term, I think both XM and SIRI will be viable, profitable companies. XM certainly is the better run, more mature of the two, but in a duopoly, there will be room for both to succeed. My only point is that prices have gotten ahead of reality, and that at least in the case of SIRI, the market is now discounting 24-30 months of sub growth.

Chris,

If calculating anything 5+ years out is foolish, especially in technology, why do you assume that anyone will be using ipods in five years instead of satellite radio?

The technologies are not mutually exclusive, indeed, they are complimentary,* but if you see it as a zero-sum game, then you cannot simulatenously say one will destroy the other while dismissing the notion that the future can be modeled.

(*A combination mp3/satellite radio device, probably on its way this year when XM’s 4G chipset is ready, will be tough to beat. Especially when listeners can hit a button when they hear something they like, and then be able to buy that song. Guess who has the biggest digital music library…)

In any event, those of us who are interested in investing and assigning value to stock prices engage in forecasting, especially in considering new technologies and immature businesses. Fatalism is not a valid method of critiquing a business plan.

You are not in the “target” group by virtue of being a 25 year old male software engineer. The “target” demographic is *everyone*. XM’s demos are broadly distributed across all age groups, and rapidly moving toward parity among gender. SIRI tries to specificaly target young men, but they probably believe they are early enough in the adoption phase to play that game, and have an unhealthy reliance on retail sales rather than OEMs right now. Still, the numbers bear out — this is a mass market product, not a niche product.

Whether the product personally appeals to you or not is irrelevant. I don’t smoke and almost none of my friends do, but I can see investing in Altria. Conversely there are many things I enjoy that would probably make irrational investments.

I’m glad XM has a good presence at the Consumer Electronics Show. Why is this a surprise? It doesn’t break the bank.

SG&A is going to be highest in the fourth quarter because of something called the holiday season, where satellite radio gets 40% of its annual net sub additions and thus total variable subsriber acquisition costs are going to be higher. But you missed the part about how each net subscriber is cash flow positive because SAC is less than the pre-pay.

Your SG&A calculations also appear to include huge non-cash expenses, so I do not understand why you hold this number up against OCF.

I’m not sure where you’re getting the accounts payable number for Q4 but XM does get alot of deferred revenue that is cash in the door, but has to be recognized over time. Most people would consider high deferred revenue a positive, because it greatly increases cash flow/working capital. (It also demonstrates consumer acceptance of the product; why pre pay if you don’t intend to keep it?)

A subscription business with fixed costs that aren’t going up, where there are no additional costs to serve each incremental subscriber, where the subscriber base is growing exponentially, and where the variable subscriber acquisition costs are less than the cash brought to the table by each new subscriber, is not what most people would call “terrible.”

The crux of your bear case appears to be no different than all the bear cases — that there is no market.

I have yet to see a bear case on XM that incorporates the reality of the market demand for the service, but shows, through an analysis of the balance sheet or other economic data, that the company is a poor investment.

I am polishing up my own notes on a rough sketch of the business in 2010 and will post those for informed, rational discussion based on numbers and reality.
——————-
Hundredyearstorm,

You are correct that per sub valuations are not an appropriate metric to judge these businesses in the abstract, as they are rapidly growing their subscriber bases.

It is, however, a useful metric, in conjunction with other operating metrics, to compare the two satellite radio companies to each other. SIRI has a radically higher per sub valuation that is not justified by differences in ARPU, churn (SIRI’s is higher), SAC/CPGA (again, SIRI’s is higher), or anything else. The difference is most easily explained by comparing the shareholder base and price per share. Off the top of my head, I believe SIRI has something approaching 400k highly excitable retail shareholders v. XM’s 2k (XM is largely owned by institutions).

Jeff raises an excellent point. There’s no reason to suppose the basic fee will remain this low. Certainly there is no resistance to it from consumers, given the rapid rate of adoption and the low churn rate. But let’s just play with the numbers we have and can deduce now.

Subscriber Base Assumption

XM has traditionally been too conservative in its subscriber guidance. For 2003, they guided to 1.2mm and came in with 1.36. For 2004, they guided to 2.8, raised to 3.1, and then came in with >3.2. For 2005, they guided to end with 5.5 for a net increase of 2.3, and already came in with net adds of 540k for Q1 (about 18% of the year’s net adds), so the trend is already pushing 6.

Guidance for 2010 is 20 million subs, although recently they’ve added the word “exceed” in calls and presentations. To me, that looks too conservative, but to bears who don’t bother plotting the growth curve or looking at OEM installation plans, it looks ridiculous. Of course, it was not very long ago when bears said XM would never reach more than 10 million subscribers.

To model this business, one must make some assumption about the number of subscribers at a certain point in time. I’ve chosen management’s guidance of 20 million in 2010, despite the fact that management has consistently low-balled guidance, and fully aware that bears who hate this business aren’t going to read any further.

I would only suggest that bears skeptical of the 20 million number please run their own analysis of subscriber size based on HARD NUMBERS — installation trends, take rates, churn. At a minimum, bears should address the historical rate of installation growth, and what is becoming known about future commitments (i.e., Hyundai will be 100% XM by 07), before providing their own number. If there is some articulable reason for asserting that the take rate will be cut in half, or the OEMs will suddenly stop installing radios, let’s hear it. Bears should also question whether they accurately predicted the growth thus far.

Subscriber Revenue

With the service costing $12.95 a month and maybe 10% at the family plan rate of $6.99 a month (additional radios 2-4 get the lower rate), ARPU would be $12.35/month. However, XM offers some pre-payment discounts, reducing the basic fee to as low as 9.99/month for prepaying 3-5 years in advance, and eyeballing their effect, I assume ARPU is $11 a month (note the price increase just became effective April 2, so historical ARPUs aren’t going to be apples-to-apples).

Note that I am NOT including ANY subscriber revenue for premium channels (such as Playboy, 2.99.month), data services like NavTraffic ($3.99/mo for existing subs or $9.99 standalone) or WX Weather, the online service ($5? a month when launched with AOL), downloading (which they’re going to enter, doubtless), or anything else. In other words, the $11 ARPU figure is obviously light.

20 million x 11 x 12 months = 2.64 billion in gross subscriber revenue.

Revenue Share and Royalties
Assume maintain the 6% royalty fee to music rights holders, 158.4mm (if the parties cannot reach agreement it would be arbitrated, I foresee no change).
Assume of the 20mm subs, 7mm are subject to GM revenue share of 40%, or $4 a month (this figure is variously estimated between 30-50% and applies only to the first $10/month), so let’s call it 336mm.

Activation Revenue
To sustain 20mm base with 1.5% monthly churn, that’s 3.6mm churn for the year, so I figure XM grosses 7mm subs in 2010. Assuming by 2010, 70% are OEM channel and do not yield activation revenue. So 30% of 7mm is 2.1mm non-OEM subs, and let’s assume they all activate online at the lower rate of $10 a piece (as opposed to the $15 activation fee charged for using the telephone). $21mm.

Advertising Revenue
In 2004, ad sales were 8.4 million on year ending subs of 3.2. What will ad sales be in 2010 on 20mm subs? Let’s be very conservative and call it $100mm.

Equipment Revenue
Assume growth in the company-owned mall kiosks and direct sales (metron). They’re just getting started in the kiosk program and this is a very cost-effective way to gain subs. 30mm (up from 11mm in 04).

Other Revenue
A trivial line item. Fees for mailing out bills to non-credit card subs, and technology/chipset licensing fees. Company is currently averse to increasing the latter, but by 2010 the hardware manufacturers may be easier to squeeze for this without impacting retail prices. Still, my model doesn’t count on it. $5mm, up from 2.8mm in 2004. (Again, assuming zero for downloading, internet service, NavTraffic, etc.)

Canada Income
It is difficult to see the CRTC not approving the Canadian venture. GM Canada is pushing hard for it, and Bitove is well connected. A decision is expected by June so we’ll know soon enough. CSR pays for all the costs of doing business up there and sends back ~1/3 of the revenue? Assume 5mm Canadian subscribers producing 780mm in revenue.

Customer Care and Billing
This item averaged 12.77/sub in 04, down from 19.07/sub in 03, and should continue to drop percipitously as people need less care (greater experience) and greater efficiencies are realized at scale. Assuming down to 5/sub by 2010 for $100mm.

Cost of Equipment
35mm

Ad Sales
Double the staff to 12mm

Satellite & Terrestrial
Unch, 40mm (note there will be an increase through decomissioning of xm-1 and xm-2 in 07 or 08, while they are running these in conjunction with xm-3 and before xm-4 replaces the co-located 1 & 2)

Broadcast & Operations
25mm

Programming & Content
45mm

R&D
30mm

G&A
30mm

Marketing
The easiest way to figure out this number is to focus on the CPGA, which I predict will hold at 80, on gross sub adds during 2010 of 7mm (see above), 560mm in variable subscriber acquisition costs and marketing efforts. This is probably too high a number, as the technology component of SAC costs should continue to decline precipitously.

GM Amortization
40mm

Depreciation
40mm (ballpark guess, two healthy satellites instead of the Boeing 702 fiasco that is coming off the books by 2008)

Long Term Debt
165mm

MLB Payment
180mm (in escrow now but recognized in 2010)

Interest Expense and Interest Income
Call it a wash. Most of the debt is due 2009. Company should have alot of cash on hand.

Other income
Rent (mostly charged to DC MPD): 3mm
———————-
Adding and subtracting all the above, net: $1.7826 billion.
Less tax rate 35%: 1.15869 billion
Assuming 310mm shares: 3.7377/share of earnings.

Now, what kind of growth multiple will the stock be sporting? Well, the assumption here is for an avg. annual subscriber growth of 2.8mm with plenty more to go at that rate.

This is obviously a somewhat rough analysis. Criticism welcome.

In regards to some sort of streaming over cellphone technology coming in the future, why would XM and Sirius not dominate this area as well? Music programming is their primary business and it would be hard to compete against them. They are already producing the channels and it would cost them less than anybody else to provide that programming via cellphones.
In the Ipod’s case, part of it’s appeal is no doubt how terrible terrestrial radio is. Ipod’s require a lot of time and care in arranging playlists, searching out new music etc. Sometimes you just want to turn the radio and hear music which matches the mood you are in. This is something that XM does very well with it’s 68 music channels. Perhaps people in their 20’s have time to organize their playlists and find new music. A lot more people just want to listen to music. I myself tire of listening to the same 3987 songs in my Ipod.

XM has repeatedly vowed at all their conferences and presentations that their 4G chipset, arriving sometime this year, will be small, cool, and efficient enough to integrate into cell phones.

It would be quite a challenge for cell phone companies to produce the content, and deliver it over their networks, at a profit. They have to offer a sufficient number of channels to design attractive niches, and then squeeze them on their networks.

The two cell competitors thus far, Sprint’s mspot and Motorola’s iradio, are each a joke compared to satellite.

Sprint wants to charge $6 for 13 channels, hardly a value proposition compared to terrestrial.

Motorola wants to charge people to download podcasts into their cellphones to take with them, not competitive with ipods or cell phones that have plain vanilla mp3 capability.

I’m reminded of the horrible 1995 Sandra Bullock movie, “The Net,” about the evils of the internet. At the beginning of the movie, Sandra is Very Modern because she logs on a dialup modem and navigates through a website to order a pizza. Ten years later, we know it’s easier to just pick up the phone. Cell phone technology is GREAT…. for cell phones. To deliver 150 channels of digital audio to an infinite number of receivers, it’s not so great.

There is no reason to believe that a WiMax-type network will be any more reliable/complete than current cell networks. Neil talks only about “the top 10 metro areas,” which doesn’t sound very useful if you don’t live in one of these… or want to drive to metro area #11.

The idea that more channels can be offered this way for the price of XM’s depreciation appears very unrealistic. XM’s content isn’t free and more content will be more expensive. (Never mind what Sirius spends for content, they are completely out of control.)

There is also the issue that if geographic reach is so limited, there aren’t enough listeners to aggregate within each niche that is offered. This is terrestrial’s big problem. Nobody wants lowest common denominator content, but then, there aren’t enough listeners in each listening area who share enough tastes. There is a national radio market for a broadway channel, reggae, prog rock, trance… but no such local radio markets.

Could something else compete with MORE content, and thus higher content cost, at a fraction of the available listeners (top ten metro areas)? Not very likely.

WiMax will compete will cable broadband, DSL, and fiber, and will likely cost about that much if not a slight premium to these services. $30 a month? (These services are not scalable. An infinite number of radios can receive the satellite signal without impacting XM’s transmission capacity. This is not true of two-way terrestrial digital networks.)

To that price, add the costs of content, plus a nice margin to make it worthwhile.

There will always be some people who have a WiMax account for their laptops, and use it on occasion to stream music. However, it is difficult to imagine a business model that profitably combines the delivery of wireless data (at the highest capacity demands), plus premium content, on a scale that can compete with satellite radio.

There is also the OEM’s interest in bandwidth. The carmakers like XM’s datapipe into their cars. The reach is throughout North America, and its completely secure. This is why GM, Honda, Nissan, and Toyota all have some claim on slices of XM bandwidth. Can some other technology supplant this?

Another factor that works against wifi type networks is how much redundant capacity would have to be built into the network. if everybody’s listening during the morning and evening rush hour that would create a very high peak demand at certain times during the day and the lower demand for the rest of the time. Whereas the cost for XM to broadcast to an additional radio is 0.