Sirius XM and Subscriber Acquisition: A Billion Dollar Question

| About: Sirius XM (SIRI)

I recently wrote an article about Sirius XM's subscriber activity and how it was an important metric that Sirius management highlighted in its quarterly financial press releases and conference calls. I would like to take a deeper look at another closely related metric that Sirius management also highlights - Subscriber Acquisition Costs or SAC.

Sirius management has stated that most cars sold in the U.S. come equipped with Sirius XM radios. During the fourth quarter conference call, Mel Karmazin said, “And our penetration level has been picking up a little bit actually, so we're in the low 60% and we don't see that changing.” See the Q&A in the Seeking Alpha transcript. The new vehicle sales are also the source of most of Sirius’ new subscribers.

Whether you buy a new Ford F-150, a GM Corvette, or an environmentally friendly Honda Civic Hybrid, the new vehicle will come with a free trial for the factory installed Sirius XM radio. The objective of offering a free trial is to get the new car buyer to experience Sirius' exclusive content, commercial-free music and crystal clear reception, and at the end of the trial, convert to a paying customer (or as Sirius calls them, “self-pay subscribers”) while the car still smells new. After all, with a service that offers so much and “costs less than $0.50 a day,” why wouldn’t each buyer become a self-pay subscriber? Unfortunately, over the past several years, at the end of the trial, more than half these listeners decided that it wasn't worth the price.

So what does it cost Sirius to acquire subscribers in this manner? Well, SAC in the May 3, 2011, press release is defined, in part, by Sirius as follows:

"(4) Subscriber acquisition cost, per gross subscriber addition (or SAC, per gross subscriber addition) is derived from subscriber acquisition costs and margins from the direct sale of radios and accessories, excluding purchase price accounting adjustments, divided by the number of gross subscriber additions for the period ..."

The SAC over the past several years has totaled more than $1.5 billion.

Subscriber Acquisition Cost

Q1 2011



2008 (Adjusted)

Total ($millions)





These are some very large numbers, and they will fluctuate, largely based on the number of vehicles equipped with radios. In one respect, SAC is a number that has shown steady improvement over the past several years and now appears to be leveling off between $55 and $60. At the $12.95 per month standard subscription rate, the SAC appears to be recovered in less than five months. Many businesses with recurring revenue models refer to this as “payback” and would be thrilled to recoup their investment so quickly. One would think that we, as investors, should be very satisfied with such a short payback period. I’m not.

Subscriber Acquisition Cost

Q1 2011






(Pro Forma)

Per Gross Subscriber Addition






I'd like to draw your attention back to the definition. SAC is the cost to acquire a Gross Subscriber Addition. As investors, shouldn’t it be important to know how much Sirius spends to acquire a new self-pay subscriber? Or even more important, how much Sirius spends to acquire a net additional self-pay subscriber? After all, it is the growth in self-pay subscribers - those willing to pay “less than $0.50 a day” - that should drive most of the increases in revenue, and ultimately earnings and free cash flow.

Trying to make meaningful year-to-year comparisons of the cost to acquire each net additional subscriber or each net additional self-pay subscriber is difficult at best. Not only have the net additional subscribers fluctuated substantially over the past three years, but the total dollar expenditures include myriad merger related and purchase price accounting adjustments. There are even timing issues and certain expenses which one could reasonably argue should be included, but are explicitly excluded from the calculation. From the 10K: “The majority of subscriber acquisition costs are incurred and expensed in advance of, or concurrent with, acquiring a subscriber. Subscriber acquisition costs do not include advertising, loyalty payments to distributors and dealers of satellite radios and revenue share payments to automakers and retailers of satellite radios.”

This tells us that it probably cost more than $456.4 million to acquire subscribers in 2010 because it could be reasonably argued that revenue share payments and the like are part of the cost of acquiring subscribers. (I have ignored the timing issue on the assumption that the difference is not material for the purposes of this exercise and should balance out over the long term.)

Next, the 10K shows that Sirius' total subscribers increased by 1,418,206 and that self-pay subscribers increased by 982,867. Using the $456.4 million subscriber acquisition cost we can calculate the following:

Acquisition Cost For Each Incremental





Self-pay Subscriber


Using the $464 figure as the acquisition cost per net additional self-pay subscriber, and the $12.95/month subscription fee, we see that it will take, on average, almost three years to recover the cost of adding each incremental self-pay subscriber in 2010. That’s quite a long time.

But, is this an accurate way to look at the payback? From a purely arithmetic standpoint, it’s a correct average. The problem with examining payback in this manner is that the net increase in self-pay subscribers (or even total subscribers) is affected by cancellations (or churn) from the entire self-pay population. There is another way to look at this cost per net additional self-pay subscriber.

Start with the Sirius SAC figure of $57 in the first quarter of 2011. Since 55% of the paid promotional subscribers never convert, it means that each self-pay subscriber acquired through the OEM channel “initially” cost $57/(1-55%) or $127. We also know that approximately 24% (the self-pay monthly churn of 2% multiplied by 12 months) will cancel during the first year. Assuming they cancel evenly over the year, the expected value of the average monthly revenue in the first year is $12.95 * (1 – ( 2% x 12 / 2) ) = $11.40. Using the $127 adjusted SAC and the $11.40 gives a payback of just over 11 months. It's a lot better than three years.

The degree of payback that can be expected to improve depends on the success of some actions and programs announced by Sirius. Internet subscriptions and smart phone applications obviously don’t require subsidizing radios. Neither do programs like CPO and promotions targeted at reactivating radios whose subscriptions have lapsed. Also, 2.0 could help with improving the conversion percentage. A price increase could also help, provided it does not result in a decrease in the conversion percentage or an increase in churn. Will it be significant?

Here’s one more set of figures to consider. According to the 10K for 2010, between Revenue Share (approximately $240 million), Sales and Marketing ($215.4 million) and Customer Service and Billing ($241.7 million), Sirius spent approximately $700 million. All these expense are not exclusively subscriber acquisition costs but they are part of the cost of growing and maintaining the installed base of Sirius subscribers. Added to the $456.4 million that Sirius actually designates as SAC, it is a number significantly above $1 billion.

As an investor, I’m satisfied when I see the subscribers grow and watch the revenue and free cash flow improve. As an investor, I look at the dollars spent on acquiring subscribers and wonder if there isn’t a better way. I wonder why Sirius doesn’t try programs similar to the cell phone services I’ve used in the past. If I wanted to get a free or discounted phone from AT&T (NYSE:T) or Verizon (NYSE:VZ), I had to sign a two-year agreement and/or submit rebate forms. If I were to cancel early, I was faced with a stiff termination charge. Is it still necessary for Sirius to offer subsidies to the OEM and revenue sharing? Is it possible to move toward a model where the satellite radio becomes a factory installed option, paid for by a buyer? What if the option were coupled with a substantial rebate for those willing to sign up for a multi-year subscription, but also had an early termination fee? Would it work?

In short, is there the possibility of developing a more cost-effective way of acquiring subscribers? As an investor, I’d like to think so.

Disclosure: I am long SIRI.

Additional disclosure: I am also long Verizon and AT&T. I have written covered calls against most of my Sirius long position. I have no positions in other companies mentioned in this article.