A few months ago, I wrote an article about Sirius XM Radio (SIRI) titled "Sirius XM Subscriber Dissatisfaction Increases". That article focused on the Q3 up-tick in the self-pay monthly churn to 2% as an indicator that subscriber satisfaction was decreasing. In the fourth quarter, the number declined to 1.8%, and Seeking Alpha contributor Stephen Faulkner took the opportunity to poke some fun at yours truly, quickly publishing an article titled "Sirius XM Subscriber Satisfaction Increases". Did subscriber satisfaction increase in 2012?
It is an important question for Sirius XM investors. Subscription revenues are what drives the success of Sirius XM. Sirius XM derived less than 2.5% of its $3.4 billion of 2012 revenue from advertising, and more than 95% of its revenues from paid promotional and self-pay subscribers. Internet radio alternative Pandora (P) is almost the exact opposite, with $106.3 million of its $120.0 million total in Q3 Revenue -- or almost 90% -- from advertising.
It's why metrics like OEM penetration, conversions and self-pay monthly churn are so important. How did each of these metrics fare in 2012, and what can we expect from 2013?
For the past several years the penetration rate in new OEM vehicles has steadily increased. The penetration rate is the percentage of new OEM vehicles that come equipped with a Sirius or XM satellite radio, with the buyer getting a free trial. The expectation is that a certain percentage of those that have had the opportunity to experience the diversity of audio entertainment for three to 12 months will become "hooked" on the service and become self-pay subscribers.
During the conference call for the fourth quarter of 2010, former CEO Mel Karmazin made the following comment:
So we believe that there is upside, not just in 2011, but in subsequent years in so far as car sales go. 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. Higher car sales, higher penetration, good conversion would get us to a different subscriber number but we don't have the confidence today to give you that number.
By now, we know that the penetration rate has steadily risen. At the UBS Global Media and Communications Conference on December 3rd, CFO David Frear presented a slide that showed the penetration rate for 2010 was 63%, and that it jumped to 68% for 2011. On the most recent conference call, CEO Jim Meyer said:
Our overall penetration rate was 68% in the fourth quarter and 67% for the full year 2012. We expect to be in this range over the next few years as OEMs remain very committed to satellite radio, including the deployment of our new 2.0 radios in additional models as well as at additional OEMs.
Our penetration new vehicle production remained at about 67% in 2012 and for the first time, new vehicle installations eclipsed the 10 million mark for the year.
And, during the Q&A, the following exchange took place:
Jim Meyer ...I personally believe, still, that high-60%, 70% is the right sweet spot for us for penetration, and I do believe that's kind of where you'll see our penetration stay over the next five years.
Benjamin Swinburne - Morgan Stanley And you wouldn't expect any change to conversion ratio as a result of that inching up over time as you get into maybe lower price point cars and stuff?
Jim Meyer It's a challenge, I'll be honest with you. I mean, we have issues of mix that we wrestle with every quarter. The primary reason is, obviously, the lower price models convert at a lower rate than the higher price models. And so we wrestle with that. I think we're still comfortable with our range that we've given for conversion, and I think we'll see how 2013 goes.(emphasis added)
Clearly, the penetration rate has increased from the low 60% level where Karmazin stated, "we don't see that changing," to the 67% level to 68% level. Now consider that with 15 million new vehicle sales expected for 2013, each 1% increase in penetration results in 150,000 additional cars equipped with satellite radios -- the penetration rate is very important. Also note the bold sentence in Meyer's last comment about conversion.
The 10K defines conversion as follows:
New Vehicle Consumer Conversion Rate is the percentage of owners and lessees of new vehicles that receive our service and convert to become self-paying subscribers after an initial promotional period. The metric excludes rental and fleet vehicles.
The conversion percentage does not include used cars. Previously, management had stated used cars convert at a lower percentage than new vehicles. And while the penetration rate has increased, the conversion percentage has, at best, held steady, and at worst, declined.
|Q3 2012||Q4 2012|
The quarterly rate has not exceeded 45.5% in the past two years, despite a modestly improving economy. This ties back to the Meyer comment that "obviously, the lower price models convert at a lower rate than the higher price models." As the penetration rate increased, more lower priced OEM models were included in the trial program.
How is the used car program faring? Well, we know that the number of dealers in the program has exploded over the past two years from 103 in Q1 2011 to more than 8,000. We also know that the used car program contributed more than 1 million gross adds, or free trials, to the subscriber funnel in 2012, and that it is projected to add 1.5 million gross adds in 2013. At what percentage will they convert?
We just don't know enough to make an informed decision because management has refused to disclose the data. When asked at the recent conference call, Frear answered:
And on the used cars, I'd say that it's still sort of early days on individual metrics, and I know it sounds like we've got a lot of transactions, and we do. But we do like to see these things sort of trend over time. Overall profitability on used cards is going to be certainly as good as the new car profitability, the single biggest reason being that we don't have to reinvest in the radio. So reacquiring revenue generating subscriptions on previously installed radios is an immensely profitable business for us. I think it will be a little while before we're able to tease out sort of sustained differences in churn profile. So you just have to stay tuned for that.
I can understand the reluctance to disclose churn. It takes a while to finish the trial, wait until after an initial subscription period has ended, accumulate and analyze the data, and maybe tweak the way the call center offers renewals and incentives. But conversions? By the time the conference call took place, the company had a sample of more than 1 million free trials to assess and clearly knows what is occurring.
There are two logical reasons not to disclose the information on conversions. The results are much better than expected, and management wants to control market expectations. Or, they are converting at much lower rates than previously anticipated.
During the Liberty Media Corporation (LMCA) 2011 Investor and Analyst Meeting on November 17th, then-Sirius CEO Mel Karmazin noted that the conversion rate for used cars was in the "mid to high 30" percent range.
If only Frear had been more forthcoming about the used car metrics...
Self-Pay Monthly Churn And ARPU
Churn was the brightest spot in the Q4 results, and Meyer stated:
The fourth quarter at 1.8% was our best quarterly performance since Q3 '08. You can really see the strength of the business in our self-pay subscriber additions...
Churn and Average Revenue Per User per month, or ARPU, are closely connected. Churn can be reduced by offering discounts to retain subscribers, and, in the past, management has stated that 1 in 7 subscribers are receiving retention discounts. Retention discounting is likely one of the reasons that 2012 subscriber additions were strong and that ARPU did not increase more significantly in 2012, despite a price increase instituted on January 1, 2012. Also working against a larger increase in ARPU was the lingering effect of a decline in the Music Recovery Fee, and the fact the price increase was gradually rolled out to self-pay subscribers as they came up for renewal.
2012 ARPU was $12.00 vs. $11.58 for 2011, and in Q4 of 2012, ARPU was $12.12, an increase over the year earlier Q4 of $11.61, but a decline from the Q3 2012 figure of $12.14. That's correct. The ARPU actually declined from Q3 to Q4, and it is not likely the result of a change in the mix between self-pay subscribers and Paid Promotional subscribers (which have a lower average revenue per month). Self-pay subscribers made up 81.9% of subscribers in Q4 vs. 81.5% in Q3.
It appears likely that the churn and subscriber growth were "managed" in Q4 by an increased offering of retention discounts. We don't know this for a fact, but read what Frear said on the conference call during the following exchange:
Bryan Kraft - Evercore Partners Question on ARPU. It was basically flat quarter-over-quarter, despite a lot of fourth quarter renewals at the new higher price point. Was there a step up in the level of discounting to drive more sub growth in the quarter, and if so, was that driven more toward retaining customers who were receiving the price increase for the first time? Or was it more to convert unpaid trial subs to paid? If you could just shed some light on that, it would be helpful.
David Frear There's no change in the discounting practices. That wasn't really a factor in the quarter on quarter change in ARPU. I think you've seen other seasonality in the ARPU in previous years, where any enhancement in is a little bit less fourth quarter from third quarter. Part of it is that it is an average, right? So the fourth quarter renewals at the higher price, remember, are coming at the end of the quarter, and there's very little recognized revenue that comes from all those rollovers. So that's certainly going to affect it. But there's really no underlying change of business practice that you should be concerned about.
Frear did not directly answer the question. He avoided saying that there was not a "a step up in the level of discounting." Instead, he framed his response in terms of "There's no change in the discounting practices." Maybe it's the lingering distrust from the recent candidate debates, where questions were rarely answered and the politicians constantly avoided direct responses. Maybe it's because Frear chose not to disclose the results of the used car program. Maybe it's just my skepticism, but I believe the churn was managed in Q4.
It's not necessarily a bad thing if churn is managed, or that ARPU declined slightly in Q4. As long as, on balance, total revenue and free cash flow continue to grow, it's a positive sign for the company. But to jump to the conclusion that subscriber satisfaction is increasing may be a bit premature.
We were told by Meyer that:
...for the first time, new vehicle installations eclipsed the 10 million mark for the year. That is nearly three times the level of installations we had in 2006.
The 10K showed record 9,617,771 "Gross subscriber additions," but it also showed a record number of "Deactivated subscribers" -- 7,610,259. "Deactivated subscribers" eclipsed the 7 million mark for the first time.
When potential subscribers get to experience satellite radio, they might enjoy the service, but on average, they don't enjoy it enough to pay for it. In 2012, 55% of new car buyers that had to make a decision when their trials expired decided not to pay for the service. Additional subscribers could be enticed to pay for it, but only after receiving retention discounts.
And if used car buyers exhibit Karmazin's projected conversion rates, an even greater percentage decline the service. If one considers that used car buyers may be likely to exhibit similar behavior to buyers of lower price point new cars, that would also support the idea of a lower conversion rate. That would point to 60%-65% of used car buyers as unwilling to pay for the service.
Equally important, whether churn is 1.8%, 1.9% or 2.0%, it means that nearly one-fifth of the subscribers that decided to pay for the service will cancel each year. And whether or not they are being offered retention discounts or paying full price, that means that 4 million self-pay subscribers need to be recruited simply to replace those that leave due to churn.
The used car channel has provided, and will continue to provide, a lift to subscriber growth. So will retention discounting. However, as long as more than half of those that continue to reject the service after receiving a free trial, and Sirius XM continues to lose one-fifth of its self-pay subscribers each year, subscriber growth will be constrained.
If ARPU is sacrificed for subscriber growth, I have a hard time accepting the fact that subscriber satisfaction is increasing. Getting subscribers to pay for the service is a challenge, and getting subscribers to pay full price for the service is a bigger challenge.
Disclosure: I am long SIRI.
Additional disclosure: I also have $3.50 January 2014 covered calls against a portion of my position, and may open new positions at any time.