TiVo By The Numbers: Reloaded

Financial analysis isn’t something I’m prepared to tackle publicly, so I’ve brought in some muscle for a multi-part series on TiVo’s numbers. Obviously this is speculative in nature and just one stockholder’s interpretation of the limited information TiVo chooses to disclose. Your mileage may vary. -DZ

Preface

One of the problems one encounters when doing research, or “digging into the numbers” is that one can always dig deeper. In digging into the later parts of this TiVo story, we discovered that we could derive much more accurate and robust figures (particularly for churn and revenue from subscriptions and advertising) than we had initially. While our general conclusions are the same, the new numbers are so much of an improvement in terms of both accuracy and reliability that we felt compelled to revise them. What follows are revised numbers, some greater detail as to the way we derived them, and some clarification (we hope) of some of the ideas presented in the earlier version. Of particular interest are Appendices A and B, which are all-new, and discuss the way we derive the important churn and revenue numbers. We have also added a new asset class to TiVo’s valuation that we did not consider before. Finally, we have dropped the discussion of ROI and MIRR as they lengthened the article and did not provide much additional insight over the “gold standard” of NPV.

As before, we try at all opportunities to make conservative assumptions. We realize that our assumptions are conservative almost to the point of pessimism, and we have been accused of just that. But TiVo is a company that lives under a great shadow of Wall Street pessimism, so any show of optimism may be seen as a lack of credibility. We leave it to the reader to decide by what factor our estimates need to be adjusted for our excesses of pessimism.

INTRODUCTION

TiVo is an enigmatic company. While management peppers us with regular press releases hyping their latest deal or newest technology, it rarely provides the kind of information we need to put a value on anything – be it a new advertising relationship, distribution deal, or their own financial statements. In this article we will engage in a bit of 8-K and10-Q exegesis in an effort to understand what is really going on at TiVo.

PART 1
The Value of TiVo’s Subscribers

First, we will take a look at the value of TiVo’s subscribers (something CFO Steve Sordello specifically declined to do in the 3QFY07 results call), and find some interesting details along the way.

To find the value of a sub, we’ll need a few pieces of information: how long does a TiVo subscriber remain a subscriber, how much does he pay, how much advertising revenue does he produce? (Note that through most of this discussion we are referring to “TiVo-owned” subscribers, and not TiVo’s DirecTV subs, as they have an economy all their own.)

Existing Subscribers Paying Recurring Fees

TiVo’s reported churn hangs in the 0.9% to 1.0% range these days, but that isn’t very helpful. We’ve done some digging into the numbers (see Appendix A for a discussion of churn), and found the churn for subscribers paying recurring fees is just below 1.4% per month. To find the lifetime of the average subscriber, we want to know how many months go by before half of a given body of subscribers has churned away. That is, we need to solve the equation:

• 0.5 = (1.0 – churn)^N,

which is:

• N = log(0.5) / log(1.0 – churn)

(Where “churn” is the fractional churn per month: in this case 0.014, i.e., 1.4%.) The result is that the average recurring-payment subscriber lasts 50 months.

In Appendix B we detail how we find both the subscription fees paid by, and the advertising revenue attributable to, each recurring subscriber. These numbers are $10.93 and $0.49, respectively. If we take a weighted average of TiVo’s “cost of service revenue” line for the last four quarters, we discover that it costs TiVo about $2.27 per subscriber per month to service the TiVo-owned subscriptions. Putting it all together, we have:

• CFPS_R = $10.93 + $0.49 – $2.27 = $9.15

(CFPS = Cash Flow Per Subscriber, the “R” subscript indicates subscribers paying recurring fees.) We know the average subscriber lasts 50 months, and if we assume a cash discount rate of 12% per year, we get:

• NPV_R = $358 per subscriber

(NPV = Net Present Value.) Note that we do not have to consider subscriber acquisition cost (SAC), because the cost of acquiring the existing subscribers has already been absorbed into the balance sheet. NPVR is the value of a recurring subscriber if you were to sell him for the cash flow he generates. There are about 893,750 of these subscribers today, so collectively they are worth $320 million.

Existing Lifetime Subscribers

Lifetime subscribers generate no subscription revenue. We will say it again: Lifetime subscribers generate no subscription revenue. TiVo gets the cash for these subscribers up front, and then amortizes it over four years (at $6.23 per month) on the income statement, moving cash from the balance sheet, through the income statement, and back to the balance sheet (with brief visits to the cash flow statement). But no money comes to the company that isn’t there after the subscriber is first acquired. (Note that we are talking about existing lifetime subs. Acquiring new lifetime subs is another story – one that we will tell shortly.)

For the purposes of the NPV of the existing subscriber base, there is absolutely no difference between a lifetime subscriber that has reached the end of the amortization period, and one who hasn’t, since both generate advertising revenue and both have costs associated with maintaining them, and neither generates subscriber revenue. The lifetime subs generate:

• CFPS_L = $0.49 – $2.27 = ($1.78)

(The “L” subscript indicates lifetime subscribers.) Yes, that’s a negative number. In Appendix A we found the lifetime of the average lifetime subscriber is 88 months. This yields:

• NPV_L = ($104),

or a total of about minus $76 million for the entire group of 731,250 lifetime subs.

This gives us a total of $244 million for TiVo’s current TiVo-owned subscriber base using a discounted cash flow valuation.

Existing DirecTiVo Subscribers

TiVo is no longer adding subscribers through its deal with DirecTV. If we assume the DTV subs will churn away at 2.2% per month (this is just a conservative guess, since we do not know the actual churn rate yet), and then completely disappear at the end of the current 3-year contract extension (again, an extremely conservative estimate, especially considering that DTV’s new management will not be nearly as anti-TiVo and pro-NDS as the previous management), and provide $0.92 per month per sub (the recent average), the NPV of these 2.8 million DirecTiVo subs is about $56 million, giving TiVo’s total sub base an NPV of $300 million.

Notes on this section:

• We have assumed that churn remains constant. If churn increases, the value of the recurring subscribers will be less than that stated, if it decreases, their value increases. The situation is reversed for lifetime subscribers (i.e., the longer they last, the more of a drain they are).
• We have assumed that advertising revenue remains constant. Since we know advertising (and other subscriber-based revenue, like viewer metrics) is an increasing business for TiVo, our valuation is underestimating the true worth of a subscriber, possibly by a substantial amount.
• Lifetime subscribers probably account for a smaller portion of the cost of service revenue, since they do not incur credit card fees, retention costs, etc. Since their cost is overestimated, their value is probably underestimated relative to the recurring subs. There is insufficient data at this time, however, to estimate the proportion of the costs to assign to each of the two subscription classes.
• TiVo’s entire advertising business is more valuable as the total number of “eyeballs” increases. Therefore the lifetime subscribers have a positive impact on the total ad revenue TiVo can generate, a factor that is not included in our DCF analysis.

PART 2
The Value of Adding a Subscriber

In Part 1, we found the value of TiVo’s existing subscriber base. But things are changing at TiVo: Lifetime subscriptions are no longer available, and new subscribers can get free hardware, though they pay more in service fees. We need to know how this affects the value of new subscribers and how much TiVo should spend to acquire them.

To examine these issues, we will continue to look at the Net Present Value (NPV) of subscribers. For the purposes of these calculations, we will use a 12% annual (1% per month) cash discount rate, as we did in Part 1. Again, keep in mind that in this analysis, we are looking at value of bringing new subscribers on board, whereas in Part 1, we looked at the current value of existing subscribers.

To get a baseline for comparison, let’s consider the acquisition of subscribers under the old pricing scheme. First, we note that in FY06 SAC averaged $196 (restated), and churn was nearly 1.6% per month. For this FY, we will be using a SAC of $260, which is our estimate for the average for the year, and a churn of about 1.4% as we found in Appendix A.

At least some of the reduction in churn this FY was likely the result of Rogers’ introduction of the minimum service commitments late last year, which undoubtedly had some negative impact on sales, and therefore increased SAC. In addition, some increase in SAC was likely the result of increased competition and general market conditions that should, for comparison purposes, be assigned to the subscribers that would be acquired if the old pricing system were in effect today. In order to provide an even comparison, we are going to split the difference on churn, giving the acquired subs under the old system a churn of 1.5% (or a sub lifetime of 46 months) instead of 1.6%, versus 1.4% (50 months) for subs acquired under the new system. We will however make the new system bear the full weight of the SAC difference (i.e., we will use $260 for SAC under the new system vs. $196 under the old system). We believe these assumptions are extremely conservative.

We found last time that existing non-lifetime subscribers generate $9.15 per month in cash flow, and have an NPV of $358. Here, however, we increase their churn to 1.5% (as discussed above), effectively shortening their life from 50 months to 46 months, and giving an NPV of $336. To find the NPV of acquiring a new subscriber under these conditions, we simply need to subtract the subscriber acquisition cost (SAC) from the cash flow. In FY06, SAC was around $196, giving:

• NPV_OLD,MONTHLY = $336 – $196 = $140

The value for acquiring new lifetime subscribers, back when that option was still available, is found in a similar manner. (We need not make subscriber life adjustments here, because lifetime subscribers are largely independent of the churn considerations we discussed above, and the option is no longer available in any event.) As we found in Part 1, existing lifetime subscribers provide an NPV of negative $104. But when they are acquired, they produce $299 in cash. Subtracting the FY06 SAC of $196, the lifetime subscriber produced $103 net cash up front. This gives:

• NPV_OLD,LIFETIME = $299 – $196 – $104 = ($1)

That return indicates that the opportunity cost of lifetime subs versus recurring subs was substantial. But because lifetime subscribers are immediately cash-positive, they provided cash for operations, so we can see, to a certain extent, the appeal.

But the fact is that there is not an unlimited demand for TiVo’s products. A lifetime subscriber at NPV ($1) took away some opportunity to acquire a monthly subscriber at an NPV of $140. While there were undoubtedly some subscribers who opted for lifetime subscriptions who would not have come to TiVo had that option not been available, some percentage would have become recurring subscribers, and the value that that percentage would have brought, was lost to the company.

So, while Tom Rogers and Steve Sordello complained during the 3Q conference call about the lifetime subscribers’ drag on ARPU, the real reason for eliminating the option was its huge opportunity cost to TiVo in terms of not getting a higher-value subscriber. One of the most important things Tom Rogers has done for TiVo is correcting this huge mistake in the company’s product offerings. It is, frankly, shocking that this financial disaster was allowed to continue as long as it did.

The NPV for an average subscriber under the old program is simply a weighted average of the monthly and lifetime subscribers, with the weights being their relative proportion of subscriber additions. When the lifetime program was eliminated, that mix was about 70% monthly and 30% lifetime in any given period.

• NPV_OLD,AVERAGE = $140 * 0.7 + ($1) * 0.3 = $98

This number tells us that even under the old system, with all of its attendant problems, TiVo was able to acquire a body of subscriptions with a substantial positive value.

Back To The Present

During the 3Q call, management told us that the average monthly subscription price under the new pricing program was about $13. If we use this value (we actually use $12.78, see Appendix B), instead of the $9.55 we used for subscribers under earlier programs, we find that these new subscribers produce $11.00 in monthly cash flow (after adding advertising revenue, and subtracting service costs). That gives us a DCF value of $431. Subtracting this year’s SAC ($260), we get:

• NPV_NEW = $431 – $260 = $171

That means the new subscribers are more that 70% ($73) more valuable than the average subscriber obtained under the old pricing program (NPV_OLD = $98), and $31 more valuable than even the recurring subs under the old system despite the large difference in SAC.

This finding helps explain something that many people found troubling about management’s statements during the 3Q conference call. Rogers and Sordello indicated that, after a year of trumpeting the virtues of the zero- or low-upfront pricing, they were going to reconsider it, and likely move to a marketing plan that involved a lower hardware subsidy and a higher advertising spend. This appeared to be another flip-flop from a company that cannot seem to stick with a plan. Less emotionally, it raises the question of whether any marketing plan can significantly improve TiVo’s subscription growth.

Looking at the options (hardware subsidy vs. advertising), we find that the subsidy is low risk – each dollar spent produces results – but costs go up linearly with results, and the approach does little to improve brand or product awareness. Advertising, on the other hand, is higher risk, since it can produce no results whatsoever. But its costs do not scale with results – costs are, in fact, essentially independent of results – and advertising improves brand and product awareness even if it does not lead to increased sales immediately. As Rogers pointed out during the call, TiVo has invested heavily in product differentiation, but in a competitive environment, that differentiation does them no good if consumers are not aware of it.

If we consider the NPV numbers we just derived for subscribers under the new and old pricing plans, we find that under the new plan, TiVo could increase SAC to $333 and have the same NPV as subscribers under the old pricing plan. Put another way, TiVo’s price increases have allowed the company to spend $73 more to acquire a subscriber. (This is on top of the $64 increase we have already seen this year.) Given annual gross subscriber additions of about 500K, that $73 per sub amounts to a $36.5 million advertising/marketing campaign – more than doubling TiVo’s entire sales and marketing budget for the past twelve months (which includes fixed costs, overhead, etc.).

A $36 million ad campaign would at least triple TiVo’s advertising spend. If this program produced any increase in sales, the costs would be spread over a larger number of acquisitions, decreasing the SAC and increasing the NPV of each. Were increased advertising successful, TiVo could cut the hardware subsidy and increase advertising further. TiVo will, of course, not spend this much (at least not immediately). But the changes the company has implemented have given management the flexibility to attempt some more significant programs in pursuit of growth. Whether or not those programs will be successful remains to be seen.

The Series 3 DVR

TiVo’s flagship product is its recently-introduced Series 3 HD DVR (the “S3”). We do not include this product in our valuation because it is simply too new (introduced during 3Q) to include in the model. But the product does illustrate a low-SAC, high-NPV approach that is inspiring, to say the least. While its high-end $799 MSRP suggests that it will not be a major driver of new subscriptions, the NPV of its subscribers is so great that it doesn’t need to generate big numbers of new subs. The box carries no rebate or hardware subsidy, and speculation (though no hard data) suggests the hardware cost is in the $400 to $500 range, giving TiVo a margin – depending on the direct-versus-retail mix – of (let’s guess) $250. Thus the NPV of a Series 3 subscriber is:

• NPV_S3 = $250 + $431 – $100 = $581

Where $431 is the discounted cash flow of a subscriber under the new pricing plan, and $100 is the non-hardware-subsidy part of this year’s SAC. $584 is more than three times the value of ordinary subsidized subscribers, and nearly six times the value of subscribers acquired last year. If TiVo can sell just 20,000 of these boxes in the fourth quarter (an attainable goal, certainly), those sales will produce the value of 68,000 ordinary (subsidized) subscribers, or nearly 120,000 of last year’s subscribers.

TiVo is also offering current lifetime subscribers a one-time option to transfer their lifetime subscription to a new Series 3 for $199. Using the same numbers, we get:

• NPV_S3LT = $250 + $199 – $100 – $99 = $250

(Where $99 is the negative cash flow over the life of a lifetime sub). That gives TiVo a sub with an NPV of $250 and, more importantly, nets the company a profit from a money-losing lifetime sub acquired under the old system.

Even at a lower MSRP, the economics of the S3 are stunningly positive. Like the luxury car lines at major automakers, the S3 gives TiVo not only cachet as the premier DVR maker, but it provides profit margins over a subsidized box that are so favorable as to warrant significant marketing attention in its own right. It is clear from this model that a companywide reduction in hardware subsidy is almost certainly in TiVo’s future.

Notes:

• While churn is an entirely negative factor, the higher NPV of new subs tends to cause us to somewhat overestimate its impact. As older, lower-value subs are replaced with newer, higher-value subs, the entire subscriber base improves in value. Thus, even if TiVo were at zero growth, with new sub additions only sufficient to cover losses from churn, the value of the business would still be growing for some time.
• Obviously, throwing advertising money into a black hole of no results is a bad thing. While the analysis above indicates that TiVo could do it from a value point of view and not lose ground to last year’s model, the results on cash flow would be extreme and painful. The truth is that increased advertising spend needs to result in higher growth to be justifiable. On the flip side of this, TiVo’s total SAC this year will be around $120 million, and that, plus the additional value of the new subs, would buy a lot of advertising.
• Our old vs. new analysis here is fairly conservative. Even if the old pricing program were still in place, SAC would likely be higher this year than last. Increasing competition from “free” cable and satellite DVRs, and the introduction of higher-priced hardware (the S2 DT) in order to remain competitive would likely have driven SAC higher. As we mentioned, the introduction of service commitments (to reduce churn) has also probably increased SAC. A higher SAC applied to the NPV of the old offerings would make them even less favorable when compared to the new offerings (which have the current SAC assigned by default.)
• In an earlier revision of this document we used $191 for FY06 SAC, however TiVo has restated some earlier costs to reflect options expenses. Since options expenses are included in the current FY’s SAC, we use the adjusted FY06 SAC ($196) as it provides for an apples-to-apples comparison.

PART 3
Finding Value

In Parts 1 and 2 we focused on TiVo’s biggest business: their subscription recording service. But TiVo has other assets, and to get a value for the company, we need to consider those items, as well. But first, we will deal with a final subscriber-related asset.

Balance Sheet and Fourth Quarter Subscriptions

TiVo’s shareholder’s equity at the end of the third quarter was about $30 million, which we could add to our valuation for the company. Since TiVo announced plans to burn through $30+ million in 4Q (which is now more than a month underway), we won’t count that $30 million as an asset. But in burning through that money, the company will acquire about 200,000 new subscribers and lose about 49,000 old ones to churn. Since the SAC is included in the quarter’s net loss, we will only consider the cash flow of the subs:

• NPV_{4Q SUBS} = 200,000 * $431 – 49,000 * $150 = $79 million

Where $150 is the weighted average NPV of existing subscribers. (We must admit, part of our motivation for showing this calculation was to demonstrate the incredible value-building power of the new subscription system. The discounted cash flow from the new subscribers is nearly three times what the average “old” subscriber generates.) Adding this number to the previous one, we get a sub value at the end of this quarter of $379 million, or $4.12 per share.

Since churn is already included in the valuation of existing subscribers, it is not really necessary to deduct them, above. However, by doing so, we move the effective date of our valuation to the end of TiVo’s 4QFY07 (i.e., January 31st, 2007).

Deferred Revenue

To understand this value, one needs to understand how TiVo accounts for lifetime subscriptions. When TiVo sold a lifetime subscription, they put an amount on the “cash” line of their balance sheet for the full amount of the subscription, and offset it with a “deferred revenue” liability of the same amount. The cash would then be amortized over the expected lifetime of the subscription. TiVo picked 48 months as the lifetime of the subscription, so for a $299 lifetime subscription, that amounts to $6.23 per month. Then, each quarter, TiVo takes the appropriate amount of cash:

• Quarterly Lifetime Cash = $6.23 * (lifetime subs – fully-amortized lifetime subs) * 3 months

from the cash line and puts it on the revenue line of the income statement, where it then makes its way through the financials like normal revenue. Then TiVo removes an amount from the deferred revenue line equal to the amount of cash shifted to the income statement.

In other words, there is a certain amount of cash on the balance sheet that is spoken for, and as it is used for its intended purpose, an offsetting liability disappears. It may seem a bit complicated, but this is a normal way to account for an upfront payment against which products or services are delivered at a later date.

So what does this mean for our valuation of TiVo? It means that when we did an NPV calculation for the lifetime subscribers that looked only at their net cash flow (excluding this $6.23 monthly phantom income), we ignored that they represented a lot liabilities on the balance sheet. (Think about it this way: the NPV we found for these subs was about negative $76 million, but if you were to sell them for this amount (i.e., pay someone this amount to take them off your hands) the “deferred revenue” liability would disappear from the balance sheet, freeing up exactly the same amount of cash.

By going back through the last four years of subscription numbers (anything beyond that is now fully amortized and therefore will no longer be represented in the deferred revenue), we have found the amount of deferred revenue attributable to the lifetime subs is $82 million (of about $108 million in total deferred revenues). Removing this liability from the balance sheet is mathematically identical to adding the same amount of equity. We therefore add $82 million to the present value of the company. This brings the value of TiVo to $461 million, or $5.01 per share. (We can add this value directly because, as we mentioned in the previous section, TiVo’s shareholder equity is about zero, so in terms of valuation of the company, the balance sheet is a wash.)

It is significant to note that this value was not produced by the lifetime subscribers. Their cash contribution to the company, as we saw, is about one-third the $299 lifetime revenue they defer, and at $6.23 per month, that cash is gone in a little over 16 months. This money, at least the majority of it, was generated by TiVo’s financing activities.

Comcast/Cox

TiVo and Comcast are a month or so away from the public rollout of their DVR offering, dubbed the “TiVo Experience” (TE) in the contract. Built upon the Motorola DVR hardware, the TE service is a premium offering that Comcast will market to draw subscribers from DirecTV (who no longer has a TiVo offering) and Echostar, as well as to motivate DVR and digital upgrades from its own customers. Cox has signed a similar contract with TiVo, with an expected product rollout in mid-CY2007. Comcast, the nation’s largest cable operator, boasts over 24 million subscribers, with over 12 million of those taking digital services. Of those, approximately 4 million take “advanced” services (defined as HDTV and/or DVR service). Digital subs and subscribers taking advanced services are growing steadily. Cox is privately held, so they do not release financial or operational data. But the company has 5.4 million subscribers, and one can expect similar averages of digital and advanced subscribers.

Most analysts put TiVo’s share of the revenue for the subscribers opting for the Comcast/TiVo option at about $1/month. While we think the true number is probably somewhat larger than this, we will use $1/month to be conservative. Similarly, we will make a fairly weak assumption that the Comcast/Cox-TiVo sub base grows to one million over the next two years, and then stays there for five more years, then disappears completely. This gives us an NPV of $46 million for these relationships.

In addition to subscribers who upgrade to the TE, TiVo is developing an advertising software platform for essentially all of Comcast’s and Cox’s DVRs, whether the consumer elects the TiVo Experience or not. TiVo will be allowed to sell national advertising on these DVRs, greatly increasing their inventory, and providing the potential for ad growth across the entire subscriber base. If we assume this platform goes live in June, 2007, onto 1.5 million Comcast boxes, growing to 4 million Comcast and Cox boxes over then next two years, then remains flat for the next five years, and we assume only an average of $0.20 revenue-share per sub per month, we get an NPV of $36 million.

This gives us an NPV of these deals of about $82 million. We note here the possibility of additional MSOs signing on with TiVo, but we do not assign a value to such possibility (as is our practice when a business opportunity is merely speculative). We also believe that our take-up rates, cumulative subscribers, life-of-contract, and ARPU estimates for these deals are extremely conservative.

Echostar Litigation

We will not rehash the entire litigation process with Echostar. TiVo has been awarded $89.7 million dollars in damages and interest, and an injunction that orders all of Echostar’s infringing DVRs (over 4 million at last count) have their ability to record television disabled. Echostar has appealed the decision, and the injunction has been stayed pending the appeal. Echostar, in their most recent 10-Q said:

“If the verdict is upheld on appeal, we would be required to pay additional amounts from August 1, 2006 until such time, if ever, as we successfully implement alternative technology. Those amounts would be approximately $5.7 million, $5.9 million and $6.0 million for August, September and October 2006, respectively, and would increase each month as the number of our DVR customers increases and as interest compounds.â€?

There are numerous possible outcomes to the case. TiVo could lose the appeal, have the case go back to the Texas Circuit court under such restrictions that they could not win, and the case is dismissed with TiVo getting nothing. On the flip side, TiVo has also appealed, seeking triple damages among other things. TiVo could win the appeal, the case could go back to Texas under such restrictions that TiVo is awarded triple damages and the injunction. In this event, the value of the case to TiVo could exceed $750 million dollars (an amount on par with the total paid-in capital over the company’s entire history). In between these extremes are much more probable outcomes, including a settlement that gives TiVo the damages it has won plus a monthly licensing fee, or a court ruling that awards TiVo its damages, but strikes down the injunction in favor of ongoing damages payments. We find the NPV of these types of arrangements to be around $300 million. Using a Bayesian approach to the various outcomes and their probabilities of occurrence, we find the value of this lawsuit today at about $275 million. This value is a probabilistic estimate of present value, and therefore does not assume either a win or a loss.

TiVo has filed no additional suits to date, however the patent in question is currently under review by the USPTO. Should the patent be validated, one could expect to see additional lawsuits filed. Cisco’s Scientific Atlanta (makers of DVR hardware for cable companies) and Time-Warner Cable seem likely next targets.

Tax Assets

At the end of FY06, TiVo had accumulated tax assets worth approximately $260 million. These assets have accrued from TiVo’s accumulated losses over its years of operations. Were TiVo to become profitable, these assets could be used to offset a like amount in taxes, greatly improving TiVo’s profitability and cash flow. In the event of a “change of control” there are restrictions on how these assets could be used, so their value to an acquiring party is unknown. Because there is no guarantee that TiVo will ever become profitable or profitable enough to use all of these assets, TiVo assigns a 100% valuation allowance to them, essentially giving them a value of $0. We will do the same, and simply note that the value is there if it can be extracted. NPV = $0.

Other Assets

There are a number of other assets for which it is hard to obtain a value:

• TiVo has launched an audience metrics business. Through their subscribers DVRs, TiVo can determine which advertisements are being watched, which are skipped, what catches a viewer’s eye, etc., all on a second-by-second basis. It is unclear how much revenue this business can generate, but in the face of $60 billion in annual TV ad spending, even small improvements in customer responsiveness are worth tens or hundreds of millions of dollars to advertisers.
• TiVo owns a 49% stake (and the option to purchase controlling interest) in TiVo Greater China (TGC). TGC markets DVR services in Taiwan and intends to expand into other Asian markets. TGC also supplies low-cost engineering services to TiVo. The value of this company is unknown, but could be significant if its distribution is successful.
• Aside from the Barton “Time Warp” patent, TiVo owns a considerable IP portfolio of over 80 patents, and many more pending applications. These patents may represent considerable potential licensing revenue.
• TiVo’s brand name is one of the most widely known and most favorably viewed in the industry. The value of such a brand in marketing efforts (such as will soon be made by Comcast and Cox) is significant, and could also be significant to organizations looking to penetrate the home media space with other products and services.
• TiVo has begun rolling out interactive features to subscribers of broadband-connected DVRs. These features include content download, games, news and information sources, and podcasts. If TiVo is able to monetize any of these features (consider features like ordering a pizza from the TV screen, renting a movie, purchasing a product, etc.), the value of their subscribers could increase dramatically.

We assign a value of $0 to these items simply because they are hard to value, and some are speculative. But together they represent a not-insignificant group of assets to which even the most cynical analyst should assign some value.

Growth

Being speculative, we assign no value to many items in the growth category. The value of additional deals with MSOs, however likely, are not estimated. Nor do we assign a value to the likely future lawsuits or IP licensing revenue from TiVo’s nearly-proven patent portfolio. We do not even account for future growth in existing business that has a history of growth, such as advertising, because of the somewhat speculative nature of projecting those trends. All of these choices make our analysis quite conservative.

One area where we can assume some future value is in subscriber additions. The DVR market is expected to grow significantly over the next several years. TiVo has demonstrated a consistent ability to add new subscribers over the course of its history, and there seems to be no signs of that ability disappearing. In fact, this year TiVo has introduced two new products (the Series 2 DT, and the Series3 DVRs) that have proven popular. We will therefore assume that TiVo is able to continue to add subscribers at the rate it did this year (FY07) going forward. To remain conservative, we will assume no improvement in SAC or cash flow for new subscribers, nor will we allow for the acceleration in subscriber growth that might come from increased advertising. We will also assume churn remains constant, which we believe is conservative because of the trend toward longer commitment terms in recent quarters (as discussed in the 3Q conference call).

Subscriber additions in FY07 were: 91K, 74K, 101K, and 200K (est.), for the first through fourth quarters sequentially. If we assume this rate of quarterly additions for the next four years, and then nothing after that (because, again, we want to be conservative), and apply our usual 12% discount rate (and no terminal value or perpetual growth rate), we get an NPV for these future subscribers of $247 million.

(In an earlier version of this analysis, we omitted this calculation. While we strive to be conservative, we found that giving no value to the actual business TiVo is in to be too much so. We feel the above value, while extremely conservative, at least acknowledges TiVo’s ability to generate value going forward.)

The Bottom Line

We have found the following:

This is a conservative estimate of TiVo’s assets today, and disregards the value of TiVo’s growth prospects and the value of many as-yet unrealized assets.

PART 4
Final Thoughts and the Future

As we listened to the 3Q earnings call from TiVo, we were struck by a number of statements made by TiVo’s management that seemed to be clues as to what we can expect from the company in the future. In this section, we will look at some of these clues in light of the financial analysis we have just completed, and see what we can learn. But be advised: we are entering an area of higher speculation and greater interpretation than we have visited before.

Subsidize Less, Advertise More

We commented in Part 2 about the company’s apparent reversal of its position on hardware subsidy after working for so long to get the boxes to a zero-upfront model:

Following the holiday period, we will be evaluating the success generated by this kind of hardware pricing approach versus an approach where there is less rebate on hardware and a greater proportion dedicated to advertising the TiVo product. (Rogers)

We noted in Part 2 the advantages and disadvantages of an advertising approach versus a subsidy approach. One particularly important advantage to advertising was noted by Rogers:

Particularly given all the differentiation that we have now worked hard to accomplish, we really think that there is a credible basis to think about advertising benefiting TiVo and not just educating people about DVRs in general, where we would not necessarily see the benefit in TiVo sales of that increased advertising spend.

Not mentioned here, but extremely important, is that increased advertising, brand awareness, and product differentiation directly benefit not only TiVo’s standalone DVR sales, but also the TiVo-branded products marketed by its partners Comcast and Cox, which are due to become available within the next few months. While these are low-ARPU subscribers, they are also near zero-cost. Any advertising spillover that benefits these programs will work to the company’s advantage. But if the company fails to adequately differentiate the products or fails to target its marketing efforts to the appropriate consumers, it will risk putting its standalone offerings in competition with these TiVo-branded integrated cable DVRs. A more general challenge will be for the company to maintain its lead in DVR technology and features, and to present those advantages to its target audience effectively and efficiently, something it has struggled with in the past.

Acquiring High-Value Subscribers

One of the most frequently touched-upon topics in the 3Q call was the “value” and “quality” of the standalone subscriber base. This relates directly to the NPV comparisons we made in Part 2, where we saw that subscribers under the new pricing program are far more valuable than those acquired under the old program.

What was also clear from our analysis was that the new subs would be even more valuable if TiVo could reduce its SAC. Nearly two-thirds of FY07’s $260 SAC will be hardware subsidy. If TiVo could moderate its growth by decreasing the subsidy, but in the process acquire a subscriber who, because he values the service enough to pay for the hardware will churn at a lower rate than subscribers who are given the hardware for “free,” the company will reduce overall costs, increase the value of subscribers, and move much closer to profitability. This is what we believe is what management is attempting to achieve.

But what is the purpose of the standalone business? More specifically, what is the purpose of growth in the standalone business, as it relates to TiVo’s overall plans? After TiVo’s relationship with DirecTV soured, the standalone business carried the weight of all of TiVo’s future. But as the Comcast and Cox relationships begin to unfold in 2007, TiVo’s advertising business will get its mass distribution far more efficiently through those relationships than through standalone subscription growth. Cox and Comcast will also provide a growing stream of high-margin revenue to replace the slowly-disappearing DirecTV revenue.

That leaves the standalone business under much less pressure. Rather than depending on the standalones to drive mass distribution, this part of the business can be used more judiciously to produce a high-margin revenue stream to fund company operations. As Rogers described it:

…what we are trying to do is come up with that right level of growth that responsibly manages our cash, comes up with the highest quality net present value of a sub, and accelerates our growth somewhat beyond what it would be if we were to run the business for profitability today.

By making the standalone business slower-growing, but self-financing, TiVo effectively un-mortgages its future. The company no longer needs push its resources to the limit to accelerate growth. The challenge here will be to find a workable balance between SAC and growth in the face of increasing competition. If reduced subsidies result in dramatically slower growth, and increased advertising does not have the desired results, TiVo will continue to struggle with high acquisition costs and may eventually struggle to even maintain its subscription base against churn. Nevertheless, the new higher-value subscribers give the company much more breathing room with these kinds of issues.

Going High-End?

One important aspect of TiVo’s business suggested by the quotes above, and others in the call, was that of product differentiation. It is clear that TiVo struggles to compete with “generic” DVRs on their own terms: cable companies distribute their boxes for no upfront cost, no commitment, and competitive monthly charges that are integrated with the customer’s cable bill. Most of the cable DVRs have dual tuners, can record and play high-definition signals, and can interface with the provider’s VOD and PPV offerings. (It is a testament to TiVo’s product and brand that despite its competitors’ advantages, the company continues to make sales in this segment.) TiVo’s best standalone growth opportunities are not in these digital cable markets where the competition comes not only from the MSOs, but soon integrated TiVo/MSO offerings.

TiVo’s best opportunities are in analog cable, where the subscribers have few other options, and with higher-end customers who are willing to pay for TiVo’s more advanced features. The Series2 and Series2 DT boxes have the analog market covered and, as we saw earlier, TiVo’s new Series3 DVR is very clearly aimed toward the high-end market. While the Series 3 box is an extreme example, it does demonstrate the desirability of attracting the “higher value” subscriber that Rogers and Sordello talked about. If, through continued product differentiation and advertising, TiVo can attract a class of subscriber that is willing to pay more for the features TiVo offers, the company stands a chance of finally stopping the flood of red ink that has plagued its history.

APPENDIX A
Churn

With the company’s quarterly reports TiVo provides an average monthly churn percentage. TiVo’s churn is the number of subscription cancellations divided by the average number of subscribers in the quarter. This number is almost entirely useless. While subscribers paying recurring fees (we will henceforth call these “recurring subs”) are considered cancelled when they stop paying fees, lifetime subs are not considered churned until they reach the end of the four-year amortization period and the box has not contacted TiVo for six months. (That means that at any given time, there are a significant number of lifetime subscriptions that aren’t even eligible to churn because they cannot be officially canceled even if the subscriber stops using the product.) Since our goal is to find the lifetime and NPV of a subscriber, we need to separate the lifetime and recurring subscribers’ respective churn rates.

The problem in tracking down these numbers is that TiVo doesn’t provide enough data to do it directly, and the data provided has changed over time. We have gone back through TiVo’s history and found the relative number of lifetime and monthly subscribers added each quarter (this can be done because TiVo has given both the total net subscriber additions and the lifetime/recurring split of the total sub base for most of its history). Four years after they are added, each of these blocks of lifetime subs becomes eligible to churn. The number of lifetime “cancellations,” then, is the number of fully-amortized subscriptions at the end of the previous period plus the number of fresh fully-amortized subs (from the four-year-ago period), minus the number of fully amortized subs at the end of the current quarter. The churn percentage is the number of cancellations divided by the average number of fully amortized subscribers in the quarter.

Even that description isn’t quite true. There are two churn numbers here: the first, a bulk number for the accumulated losses during the four-year amortization period that immediately churn when they hit the end of the amortization period, and the second churn number that is the ordinary churn rate of the fully-amortized subs. We separated these numbers as best we could by minimizing the variation in ordinary churn. The results were a 19% immediate churn, and a 1.2% per month ordinary churn. Using these numbers, the lifetime of a lifetime subscriber is given by:

• 0.5 = (100% – Immediate Churn) * (100% – Monthly Churn) ^{N – Amortization Period}

i.e.,

• 0.5 = (1.0 – 0.19) * (1.0 – 0.012)^{N – 48}

giving,

• N_Lifetime = 48 + log(0.5/(1.0 – 0.19))/log(1.0 – 0.012) = 88 months

We now move on to finding the churn of recurring subs. TiVo has provided the number of quarterly subscription cancellations since 3QFY04. From our work above, we know the number of lifetime subscription cancellations each quarter. From these numbers it is a simple matter to compute the number of recurring sub cancellations, and from that, the churn. We find that over the past four quarters recurring churn has averaged 1.38%.

(TiVo did, briefly, provide a lifetime/recurring churn breakdown: in 2Q and 3Q FY06, and for the full-year FY06. TiVo indicated the amount of the churn attributable to each of the subscriber classes for each period. Our numbers match TiVo’s almost exactly for those periods, providing us with a certain level of comfort in using our numbers for other periods.)

Interestingly, in FY06, the churn on recurring subs jumped to nearly 1.6% per month. In the third quarter of FY06, Tom Rogers, the then-new CEO, instituted a policy of minimum one-year service commitment on new subscriptions. In the four quarters since that action, churn has dropped to just under 1.4%. While it may be too early to reach a final conclusion, the service commitment policy seems to have worked quite well. That two-tenths of a percent reduction translates to the average subscriber lasting 6 months longer – which works out to $40 to $50 in NPV.

APPENDIX B
Subscriber Fees and Advertising Revenue

TiVo regularly provides ARPU (Average Revenue Per Unit (i.e., subscriber), but like TiVo’s churn number, it is almost entirely useless. TiVo’s ARPU combines lifetime and recurring subscribers, as well as non-paying (i.e., fully-amortized) lifetime subscribers, plus advertising and other subscriber-related revenue. But we want to separate out these components so we can find the value of the average subscriber in each of the subscriber classes (because, as we learned in Appendix A, they churn at different rates and will therefore have different values).

In the 3Q results call, TiVo management stated that the average monthly subscription fee paid by new subs for the latest quarter was about $13 per month, compared to about $9.50 in the same quarter last year. Normally we would be skeptical of such numbers and what they might be including or excluding, but both the CEO and the CFO mentioned these figures a total of six times, explaining the situation in some detail. The CEO mentioned it again at an investors’ conference a few days later. Said Rogers:

…we have also been able to look at the monthly price that a new subscriber is coming in and paying, and we were looking at before what was in the $9.50 range to something that is about $13 a month in terms of that average price that a new subscriber is now coming in and paying on a monthly basis…

For the purposes of this analysis, which tries to be conservative when estimating inexact numbers, we will assume that “in the $9.50 range” means “$9.55,” and “about $13” means “$12.78,” with the knowledge that we won’t be wrong by too much in either direction. (We bump $9.50 up, and $13.00 down, because Rogers was trying to emphasize the improvement in sub fees, so we assume he rounded in his own favor.)

Last year, TiVo offered three different pricing levels that must be combined to arrive at the average monthly fee: 1) lifetime subscriptions, which amortize at $6.23 per month, 2) the normal $12.95 monthly rate, and 3) a $6.95 per month multi-subscription discount (MSD) for people who have more than one TiVo box on their account. In last year’s Q3 (the one which Rogers referenced, TiVo tells us that about 74% of new subs elected to pay recurring fees, while 26% elected lifetime. The only missing value is the fraction of new subs taking the MSD:

• $9.55 = 0.26 * $6.23 + Frac_MSD * $6.95 + (0.74 – Frac_MSD) * $12.95

which gives

• Frac_MSD = 0.275

So 27.5% of all new subscribers were electing the MSD, which amounts to 37% of new recurring subs electing MSD and the remaining 63% paying the full $12.95.

If we assume these numbers are representative (which is a major assumption, but the only reasonable one to make with the information available), we can derive two important pieces of information from them. The first is the average sub fee paid by a recurring sub:

• SubFee_{R AVG} = 0.63 * $12.95 + 0.37 * $6.95 = $10.93

The second piece of information we can derive by assuming the MSD percentage is representative (and therefore, over the life of the MSD program has become representative of the entire body of recurring subscribers) is the advertising revenue in any given period:

• R_A = R_T – R_L – R_R

Where RA is the monthly advertising and audience-measurement revenue for the period, RT is the average monthly TiVo-owned service revenue for the period (provided in the forms 10-Q/K), RL is the monthly revenue from lifetime subscriptions, which is $6.23 times the average number of non-fully-amortized lifetime subs (i.e., the average number of monthly lifetime subscribers minus the average number of fully-amortized lifetime subs), and RR is the average monthly revenue from recurring subs (i.e., $10.93 times the average number of recurring subs). (This is an approximation – the technique we use is a bit more sophisticated in that it accounts for the growth of MSD subs within the overall subscriber base since the MSD was originally offered.) Averaging over the past four quarters, we get advertising revenue of $0.49 per subscriber per month. We should advise, however, that this number is sensitive to small changes in the MSD percentage within the existing subscriber base, and that is a number we have estimated from just a couple of data points. While the result we seek (i.e., cash flow from subscribers) is fairly insensitive to this number (because here we are simply apportioning the total subscriber revenue to the different subscriber classes), the absolute value of this number could be off by a considerable amount.

We also note that while Mr. Rogers and Mr. Sordello attributed the jump in fees paid by new subscribers to the elimination of the lifetime service option and to the higher subscriber fees under the new bundled plans, it is impossible to get from the “$9.50 range” to “about $13” with those factors alone (given that approximately 50% of new subs are said to be electing the 3-year service plans). There must have been a reduction in the number of subscribers opting for the MSD, as well. Since the new bundled service plans available online are not eligible for the MSD, and marketing has been targeted to new subscribers, it is not surprising that there would be a drop in these subscriptions.