The Revolution Will be Televised: Assessing Intel’s TV Strategy
Intel is hoping to disrupt the Cable industry and generate signifcant returns in the process but will it be enough to sate investors hungry for growth?
It appears the long-rumoured entrance of Intel into the cable TV business is imminent. I’ve been particularly interested in this development for its sheer absurdity; doesn’t Intel make computer chips? Somewhere on my flight between Mexico City and Toronto I decided to do a sanity check on the proposed move by analyzing Intel’s business case. For those that don’t know, Intel is reportedly planning to package the most popular channels from cable companies and sell the right to live-stream those channels over the internet. The service, sometimes referred to as “Over The Top” or OTT because it goes over the top of traditional content distributors, is expected to go live at the end of 2013.
Before we dig in to the economics let’s explore two aspects of Intel’s proposed move:
1. Industry Context
The entertainment business doesn’t like competition. As Intel prepares to enter the content distribution business existing operators are putting up creative, legally questionable obstacles. TWC, for example, is reportedly forcing cable channels (i.e. content creators) to refuse to sell to cable operators (i.e. Intel) who don’t own their own infrastructure. Though the language is likely more obscure than that the implication is clear: don’t sell to internet-only cable providers. Content producers are set to enjoy the benefits of increased demand through retrans fees – the per subscriber fee they charge to cable operators (i.e. TWC) for the right to broadcast their channels. As an aside, to understand why retrans fees are important read up on the current dispute between TWC and CBS. Getting back to the point, retrans fees are set to continue increasing rapidly. They have already increased a whopping 3,437 % between 2006 and 2012 and analysts forecast that these costs will nearly double between 2013 and 2018. What this means it that more players (read: Google, Microsoft, Sony) are expected to be competing for original content.
2. Virtual Cable Operator (VCO)
The concept of a VCO isn’t new – in the telco space alone there are numerous examples of virtual operators, known as Mobile Virtual Network Operators (MVNOs) selling services to end consumers while piggybacking off of infrastructure owned by someone else. MVNOs sell telecommunication services to end-consumers but “rent” the pipes and poles that deliver the services from larger telcos such as AT&T. To compensate the infrastructure owners MVNOs pay a per-subscriber license fee. The unique aspect is that, unlike MVNOs, Intel won’t be paying license fees to the internet service providers (ISPs) for what will likely be serious usage. Netflix, for example, already accounts for 32 % of fixed network internet traffic and as services such as Intel’s become available this type of usage will increase, potentially causing capacity issues with ISPs. The implication here is that Intel’s ability to scale will be dependent on the goodwill of ISPs, many of whom are also potential competitors. Commenting on a potential massive switch to internet cable companies from traditional cable companies Glenn Britt, the CEO of TWC, said “…[W]e would have to recover more money from the Internet service” by raising prices. Another potential option for Intel is to go the Netflix route and build out its own version of the Open Connect Network, Netflix’s proprietary content delivery network. The cable value chain has two major components: content providers and content distributers. Today, components are unlikely to be owned by the same company. However, as Netflix has shown, this may change as the industry evolves.
Now, onto the economics of Intel’s proposed business model. No one is sure exactly how Intel plans to roll out so let’s make some key assumptions.
Based on several reports I’m going to assume Intel will roll out their service to reach the largest number of potential customers in the fewest states. This assumption has two components: going where the most number of subscribers are and going where the most number of available subscribers are (by ‘available’ I mean not currently captured by the top 14 cable operators). In the US, 21 states account for about 80 % (90M) of the 112 million TV households. In addition, these same 21 states account for 73 % of available subscribers.
Seems like a good place to start, right?
Let’s call these 21 key states Intel’s “Wave 1”. Since the firm wants to rollout nationwide I’ve identified the “Wave 2” and “Wave 3” phases as well. Note that the data only covers the top 210 TV markets in the US, covering 46 states.
Narrowing in on the numbers we see that as of Q1 2013, the total available subscriber number in Wave 1 is 12 million, small fries in comparison to the other 100 M households.
Not only does Intel lack experience packaging and reselling content but there is no real pricing precedent to follow in this market. This introduces an element of complexity to our thinking but one that can be solved by looking at the extremes (i.e. what is the maximum Intel can charge?). I used the pricing from 10 major cable TV providers across the US to arrive at an average monthly price of $ 48.29. A comment from the lead of Intel’s TV business, Erik Huggers, provides a bit of guidance on what Intel’s potential pricing strategy could be: “this isn’t a service for cord-cutters or anyone who wants a cut-rate cable package. Rather, it’s a better cable experience that is designed for (and will be marketed to) the kind of young, affluent and connected consumer who would like a TV service that works as well as their tablet or iPhone.”
I’m a cynic so for now let’s assume that Intel plans to provide no value-add services on top of its TV offering. Doing so leaves us with the conclusion that Intel couldn’t necessarily charge higher than the average for its service so we can use $ 48.29 / month as Intel’s presumed pricing. The idea here is that the willingness consumers have to pay for the service based on the value-add they find from having access to TV over the internet will be limited by the price of $ 48.29.
The main input for COGS is per sub retrans fees, which are projected to grow rapidly over the next ten years. As noted above, TWC and other operators are making it difficult for content providers to do business with Intel. To counter this, Intel is breaking open its wallet and reportedly offering between 50 and 75 % in premiums over current retrans fees. I used a 75 % premium on existing retrans fees when projecting Intel’s COGS and to complete the scenario I used ratios from TWC and Comcast to simulate what a potential cost structure would look like for Intel. I understand the limitations in doing so, namely that a cable operator that owns the physical infrastructure would by definition look very different than a VCO. However, for our purposes we have no better data.
Before getting in to the math let’s explore why someone would be interested in purchasing Intel’s TV offering. In response to a July 24th court ruling in favour of Dish Network Glynn S. Lunney, a Tulane University law school professor, said “[a]s copying and distribution technologies have gone digital, consumers, not the content owners, are in charge of where and how they experience content.” I think this hits the mark in that consumers are increasingly mobile and are demanding their content go with them. Live streaming TV allows consumers to do this making it a major part of the value proposition.
In addition, there are reports from the Wall Street Journal (July 29) and Business Insider (April 4) hinting that people who have been privy to Intel’s beta TV offering have been been very impressed – “the TV service you’ve always wanted”, one person commented. Arguably the most interesting feature that Intel is preparing is an always-on DVR that would record everything on TV and store it for 3 days in the cloud. You, the consumer, would then leverage Intel’s “gorgeous interface” and some presumed Netflix-like recommendation engine to not only organize the shows you know you want to watch but also bring in additional content you wouldn’t have otherwised watched. At the very least, this should pique your curiousity.
In projecting subscriber uptake I used the basic formula in the industry:
1. Addressable Households Passed (HHP) x Market Share (%) = Gross Subscribers
2. Churn Rate (% per Year) x Gross Subscribers = Churned Subscribers
3. Gross Subscribers – Churned Subscribers = Net Subscribers
To arrive at a specific set of subscriber numbers I used a simplified Monte Carlo simulation with upper and lower bounds for market share and churn.
For market share I thought that for the first three years of market entry Intel would gain between 0 and 3 % of any market they were in. Beyond the first three years I assumed a 0 to 5 % range but because Wave 2 and 3 were going live during this time Intel only breaks 3 % overall by 2022.
My view on churn is that Intel, lacking any distinct consumer-facing support capabilities, will struggle throughout the ten year study period. Since many cable firms have stopped reporting churn I used an analyst report to arrive at lower and upper bounds of 27 % and 38 %. I assume that the first year will be rough, with low market share and high churn. In 2015 I assume Intel works out the kinks and begins to reach critical subscriber mass. I’ve deliberately modeled non-linear growth to be conservative and to account for the likelihood of the traditional cable industry retaliating in some way to Intel and other OTT providers.
I believe Intel’s TV Offering will become a billion dollar business by 2017 and reach 2 billion by the end of 2023. However, over this time period Intel’s COGS will grow much faster than revenue thus compressing net profit per sub (the red line). Based on market entry timing, changes in market share, and alternating churn rates the projection isn’t linear. This isn’t a concern given how little concrete data is available in addition to how mysterious Intel’s current thinking is.
Looking at these numbers from the perspective of Intel’s entire business we see that even by 2023 the Earnings Per Share (EPS) uplift this initiative could provide is $ 0.05 / share. You read that correctly – 5 cents per share. The $ 259 million in projected net profit in that same year is equivalent to Intel improving its current revenue base by 0.5 %. Clearly, the way we’ve structured the business makes it a very small part of a much larger whole. What I take away from this is that, from an investor perspective, Intel needs to go big or go home. Perhaps my 3 wave strategy needs to be reduced to an accelerated 2 wave strategy in order to achieve a more meaningful EPS uplift. Whatever lever Intel decides to use it needs to have significant magnitude (Pop Pop!). Of course this analysis considers neither the capital necessary to launch this initiative nor the free cash flow implications. However, we could spend an entire article exploring those topics so I left them out.
Despite the constraints we face around data we can still identify at what point the business model doesn’t make sense. Though we can alter numerous variables and create data table after data table I think the most salient variable to play with is price. It is, after all, the most consumer-facing and the one most likely to be top of mind for Mr. Huggers, the lead for Intel’s TV Offering. Looking at price, ceteris paribus, we find that Intel’s “margin of safety” – the monthly price above breakeven that Intel charges – declines 40 % over ten years from $ 10 / month to $ 6 / month. The main causes of this decline are slow pricing growth and rapid cost growth. The two forces combine to leave Intel with a maximum cushion of 20 % and a minimum cushion of 12 %. These may seem like robust margins of safety but given how often cable companies discount prices this can be a concern. Moreover, with new services such as Aereo, who recently won an important judicial victory in New York, rapidly being brought to market and other, similar competitors at the gate the issue of pricing is a key one. For my own curiosity, I looked at a similar analysis for retrans fees and found that Intel could pay between 2 and 5 times its current per-sub retrans fee and still break even. That means Intel could pay between 147 % and 402 % in premiums over current per sub retrans fees. I still believe that price is the main variable to be concerned with, especially considering that the view of most consumers can be summed up in a statement by Netflix’s Chief Content Officer: “television is television, no matter what pipe brings it to the screen”.
On July 24th Google announced a new service, Chromecast, that could potentially be a beachhead for a larger OTT offering. Chromecast allows users to wirelessly stream content from services such as Pandora, Netflix, and YouTube onto their TV. This could potentially act as a base from which Google builds out a more complete OTT offering. My key takeaway from this is that Google has set a pricing precedent of $ 35 / month, below what we noted was the breakeven price / month for a more traditional OTT service (I know Google’s price is one-time but numbers tend to stick in people’s heads). This isn’t a good sign for Intel – they should be worried about their pricing power relative to Google and their potential OpEx costs to ensure that their offering reaches net profit as soon as possible.
Another major player that has been looking into OTT is Apple, who is rumoured to be collaborating with cable operators to fix the user interface issues with traditional set-top boxes before creating its own branded TVs. Other players include Microsoft and Sony though I don’t know much about their plans. In addition, from a strategic perspective, ISPs (i.e. TWC and Comcast) could potentially create their own OTT service since it would provide a business case for expanding their internet infrastructure.
The implication here is threefold.
First, I believe Intel’s saving grace will be putting out a full-fledged OTT service before any other major player. Apple, Google, Microsoft, and Sony all have existing talent bases in both consumer-facing businesses and content-management businesses giving them an inherent advantage over Intel. Second, I predict that new entrants are likely to use sales pricing to capture subscribers initially, convincing subscribers to remain with the service by dint of their unique features. Google is already doing this by offering 3 months of free Netflix membership for each Chromecast purchased (though it appears Google has pulled the offer after 1 day – maybe they should have looked at the economics first!). What this means is that Intel, who already has limited room for pricing flexibility, will need to avoid sales pricing in its go-to-market strategy. Third, Intel will need to adjust their product development process to adapt to a market that will likely be changing on a weekly, if not daily, basis. The firm is used to product lifecycles that can have its chip remain in the market for years before requiring a change. With potential competitors such as Google and Apple, Intel can’t afford the luxury of not being able to adjust very, very quickly.
At the beginning of this article I set out to analyze Intel’s proposed business model and in doing so have found that the business can be profitable. However, after looking at the larger context we realized that while the TV Offering may be accretive it may not be the growth engine investors hope it will be. Based on this information I believe the firm’s strategy is something besides simply offering an OTT service. Even generating a return is open for debate given the lack of pricing flexibility introduced by Google’s Chromecast. Mr. Britt, who recently announced that he will be resigning as CEO of TWC at the end of 2013, echoed this sentiment when asked about Intel or Google’s ability to make money using an OTT service: “It’s not clear to me that you could make very much money going that route, and maybe not any money.” My analysis has revealed that there is money to be made in OTT despite Intel’s path forward being ambiguous and the firm facing significant complexity in implementation.
If all of the above is true, and assuming our numbers aren’t horribly off the mark, then the question becomes ‘why should Intel in specific pursue this initiative’? One potential answer is so that Intel can gain a firm footing in the mind of the consumer. Most devices (smart phones & tablets) are used to consume content and if Intel is the one packaging that content it will naturally assume a position of prominence in the consumer’s mind. Once again, the question becomes ‘why’? Intel isn’t a consumer-facing brand, at least not the same way that Apple and Microsoft are, so why would they care about being so important to consumers? I have to admit I don’t have a clear answer but I can’t wait to watch the revolution play out on my iPad.