Businesses are now selling subscriptions for software as a service (SaaS), but also for products like car washes and movies. Having a more predictable recurring revenue stream is often cited as a benefit of a subscription business model. If you look deeper there are more important benefits of an ongoing subscription relationship with customers that can also produce significantly higher return on investment, but only if the right conditions exist. There are big new challenges with a subscription business model. Whether value is being created with a subscription is not as easy to determine as when a business is selling one unit at a time (e.g. as in a traditional car wash business model). Perhaps the most challenging is that in a subscription business the higher customer acquisition cost (CAC) is paid up front and that cash goes out the door immediately. Whether a business likes it or not, it must play the game on the field and that game is now often a subscription business model. If business A is selling subscriptions successfully and business B cannot do so, the later business may be walking dead. Or not. The outcome depends on many factors that I will discuss below.
Some of the factors that might push a business to adopt a subscription business model can be illustrated by looking at an example. The numbers in the columns below are from the operator of Regal Cinemas (which was recently acquired by the British theater chain Cineworld for $3.6 billion). It is easy to see after looking at the numbers below that the business model for a movie theater is about selling complementary goods (concessions) associated with a loss leader (showing movies). In this example, the cost of concessions for the movie theater is only ~13% of revenue. In contrast, the cost of movie rental and advertising is 52% of revenue (Ouch!). For the movie theater owner, generating profitability is about the sodas, popcorn, candy and any other concessions they can sell. Innovation by movie theaters tends to be focused on the snack bar and not the movie experience. As an example, to improve the unit economics of the movie exhibition business, Cinemark has added liquor to at least 70 of its theaters and AMC has done so at about 250.
Key parts of Regal Cinemas’ financials looked like this in a recent filing (for the quarter ending September 30, 2017).
Running a business like Regal’s movie theater operation that only has a 3.1% incme from operations is, as the characters sing in the musical Annie, “a hard knock life.” Why is the profit margin so low? A better question might be: Why isn’t it negative? The answer is probably that the movie owners know that if they raise the wholesale cost of film rentals much more the theaters will go bankrupt, which would kill their ability to distribute movies in this theater format. The cost of movie rentals for a theater is made just high enough by the movie studios that the movie theaters make a thin profit margin.
What theaters pay for movie rentals isn’t uniform, can vary from chain to chain/movie to movie and is complex. The limits of what a movie studio can charge are always being tested. For example, Bloomberg reports that “Disney asked some theaters for a greater share of ticket sales for big films like Star Wars, according to a person familiar with the situation. Usually the revenue is evenly split but on very popular films, Disney can claim more than 60 percent.” Another news report indicates that Disney requires a 65% cut of The Last Jedi’s domestic ticket sales and a four-week run in each theater’s largest auditorium. If a theater operator breaks the contract with Disney (e.g., taking the film off-screen too early) there is a 5% penalty.
Logically, there is little point in a business like a movie studio killing its indirect distribution. But many things in life and business are not logical. If the movie theaters ever make a bigger profit, even from concessions, the studios can just raise the price of film rental. The movie theaters have this pricing power since the best movies are not just “substitutable” commodities. For example, Disney is the only supplier of The Last Jedi and has massive “wholesale transfer pricing power” with respect to that movie. What is wholesale transfer pricing power? I have written a blog post specifically on that but it is perhaps best explained by example:
People who do not have what Roger Fisher called a “BATNA” (Best Alternative to a Negotiated Agreement) in his book Getting to Yes have lousy negotiating alternatives due to wholesale transfer pricing. Wholesale transfer pricing is not about moats, but rather about the red “Bargaining Power of Suppliers” category in Michael Porter’s “Five Forces” model depicted below:
The “loss-leader with ugly wholesale transfer pricing” dynamic that exists in the movie theater industry results in some predictable behavior:
One theater chain executive went so far as to describe the cup holder mounted on each seat, which allows customers to park their soda while returning to the concession stand for more popcorn, as ‘the most important technological innovation since sound.’ He also credited the extra salt added into the buttery topping on popcorn as the “secret” to extending the popcorn-soda-popcorn cycle throughout the movie. For this type of business, theater owners don’t benefit from movies with gripping or complex plots, since that would keep potential popcorn customers in their seats. …theater owners prefer movies whose length does not exceed 128 minutes. If a movie runs longer than that, and the theater owners do not want to sacrifice their on-screen advertising time, they will reduce the number of their evening audience ‘turns” or showings from three to two, which means that 33 percent fewer people pass their popcorn stands.
What does the top line financial picture in movie industry look like in general? Not very good says The Economist magazine in a recent issue:
“Americans are on track to have bought around 3.6 movie tickets per person by the end of the year, down by 30% from 5.1 in 2002. They pay $8.93 for a ticket, 54% more than 15 years ago, which means, for now, higher total takings, but attendance is expected to decline further. Frequent filmgoers—those who go once a month or more—have dwindled, from 28% of North Americans in 2002 to 11% in 2016, according to the Motion Picture Association of America.”
A new business known as “MoviePass” has appeared relatively recently with a new business model with new unit economics. A NPR report describes the MoviePass product:
KULAS: When you’re at the theater, you pull out the MoviePass app.
SPIKES: So go through and pick a movie that you’d want to see.
KULAS: Then I walk over to the ticket kiosk, choose a ticket … to pay for it, I swipe with the MoviePass debit card. The normal price to see “Spider-Man: Homecoming” here is $16.29. MoviePass will pay that full price of my ticket to AMC. But my price is $9.95. And that’s not just for one movie. That’s per month. And it lets me see a movie every 24 hours. Movie theaters use loyalty programs to get data on their customers. And now some of those customers are MoviePass customers first. And that data is a really big part of MoviePass’ future plans. When you go to a movie, that’s something a bunch of businesses want to know about – maybe the restaurant next door, maybe the studio that will sell you the movie soundtrack later.
The key hypothesis in the MoviePass business plan is described as follows in a Forbes article linked to in the Notes:
“…after an initial burst of activity in which subscribers may attend four or five movies per month for a few months, the vast majority will settle into a pattern of seeing just enough movies to justify the subscription price.” We found that on a $40 per month plan, subscribers would attend an average of 3.8 times per month,” the CEO Lowe said. ‘At a higher price they would attend more frequently, and at a lower price, less. At $9.95 per month we expect the average subscriber to settle into a pattern of just over a movie ticket per month.’ With an average cost of around $9.00 per ticket, that would put MoviePass at breakeven. And even one ticket per month would be good news for theater owners, because MoviePass has shown that it can dramatically expand the theater-going audience beyond frequent moviegoers. ‘One way to look at what we do is that we provide reluctant moviegoers with a sort of ‘bad movie insurance’,” Lowe says. “We take away the hesitation and risk around buying tickets for movies that might be disappointing. People know if they see a bad movie, they don’t feel like they’ve misspent the money because they can always just go to another movie. MoviePass can also benefit from all the data it collects on its subscribers, their locations, habits and tastes. It can package and sell that data to movie distributors and marketers, but more to the point, it can also use that information to create its own internal revenue streams.’”
People paying for 12 movies a year rather than just 3.6 as the Economist noted could be good for the movie industry if the impact on return on investment is positive. But at what cost? What is the return on investment? Even if the financial impact on the industry is positive, which business or businesses in the value chain capture that incremental financial return?
Given the experience the movie theaters and the movie studios have had with Netflix, they are no doubt concerned that they may be losing an important part of the value chain to a company like MoviePass. They seem to understand that a business that owns the customer data is in a position to make a better product, sell the product more effectively and capture more value. It already makes movie studios unhappy that Netflix does not share its movie viewing data with them. Some people in the movie studio and theater business seem to think that MoviePass in the value chain would be another lost opportunity to capture valuable data. MoviePass is trying to change that view, but this will not be easy. Some people like to say: “data is the new oil.” Owning data about customers can be very valuable since it can be potentially converted into more cash flow and return on investment. In other words, having customer data has optionality, even if you do not know how to monetize that data yet. But monetizing data successfully requires a reasonable cost of acquisition, storage, processing and other factors to be present.
One risk for a business like MoviePass is that the movie studios and/or movie theaters borrow the MoviePass subscription pricing idea and do it themselves. For example, Cinemark launched a monthly subscription program called Movie Club. Members of the Cinemark system pay $8.99 a month and receive a credit for one movie ticket a month. Subscribers can also buy additional tickets for $8.99 each and get 20% off on concessions. A movie theater doing this successfully is tricky given the power the movie studios have over pricing and the complex and non-uniform pricing structure of film rentals. The Cinemark discounted price offer is very unlikely to be the last response by movie exhibitors to MoviePass. More subscription style price offerings are inevitable or at least the arrival of new methods of creating an ongoing relationship with people who see movies in theaters is inevitable.
How can MoviePass studios protect themselves against competition from movie theaters and studios? Netflix moved to protect itself from this by making its own original content. It is hard to imagine MoviePass doing the same thing. MoviePass’ ticket purchases represented 1.78% of domestic box office for Warner Brothers’ Justice League on Opening Weekend (Release Date: November 17, 2017). MoviePass claims it is responsible for an average of 2% of box office revenue nationwide. It is hard to see MoviePass creating its own movies or building is own theaters. MoviePass would benefit from having state of the art data science capability. A business like StitchFix has more than 90 data scientists. How many data scientists work at MoviePass?
Effectively using modern data science approaches is important in any business today, but this is especially true in a business like movies. In my previous blog post on the movie The Princess Bride, I quote the writer William Goldman as saying “no one knows anything” in Hollywood. That may be true, but your odds of being right are better if you have great data about your customer and are able to implement a modern data science based approaches. To illustrate the approach here is Jeff Bezos describing Amazon’s data science-based approach in a Wall Street Journal interview in February of 2000:
Since Bezos said this about A/B testing in 2000, techniques and approaches have become vastly more accurate, substantially more revealing and massively cheaper to do.
With a subscription business model the relationship between the business and the customer is ongoing and in most cases “connected” in nature, since it is online. The subscription also creates a reason for the customers to share certain personal data that can be used to create a better product. That online connection creates a stream of data about the customer’s preferences and behavior called “telemetry.” That telemetry can be used to optimize each aspect of the relationship and customer experience which improves customer lifetime value.
Business do not need to guess anymore. Selling subscriptions is not new. What is new is that once a business has data from customer telemetry they have the ability to run many experiments relatively cheaply and quickly to optimize each variable. With telemetry a business no longer needs to guess about what customers want. By optimizing each variable the business creates more value for customers which in turn improves return on investment. Businesses do not do adopt this approach to doing business are toast. The race to get an online relationship with every customer is reaching a fever pitch, enabled by increasingly inexpensive cloud computing and communications and radio technology. If you have use a search engine, or Netflix, have a connected speaker in your home featuring a technology like Alexa or Cortana you are part of this phenomenon.
As an example, mobile operators have been able to do what I am talking about for decades now since the nature of what they sold was a connection. What is new is that more and more types of businesses finally have what mobile operators have been taking advantage of for years. The good news is that this new connected relationship combined with low cost cloud computing resources and modern data science tools means there has never been so much opportunity to add new customer value and create higher service provider return on investment through creative use of telemetry than is the case right now. The bad news is that this capability can eventually become a commodity and competition means that only consumers benefit from the improvement. That is as it should be.
How is this phenomenon changing the nature of distribution? Not too long ago a business could make a product and have indirect distributors between it and the customer. For example, Disney could be very happy having movie theaters or Comcast or Netflix between it and the end customer. The business which created the product could relegate that distributor like a movie theater to terrible margins or not depending on relative bargaining power. In any event, the optimal distribution model can shift in a binary way once the creator or owner of the product needs telemetry data from the end customer. For example, Disney has discovered that it must have a direct relationship with the end customers to generate the needed telemetry. In other words, going direct to the customer without an intermediate distributor has become essential in some businesses. Businesses have “seen the movie” about the value of Amazon and Netflix’s telemetry data combined with data science based advantages and have in some cases decided that they must match this direct to consumer capability or die.
What about the potential for increased return on investment in a subscription business model I talked about above? Borrowing from a great blog post from Bill Gurley, a business starts with this simplified formula which spells out the “unit economics” of a business:
The key statistics are as follows:
ARPU (average revenue per user)to C
Avg. Cust. Lifetime, n (This is the inverse of the churn, n=1/[annual churn])
WACC (weighted average cost of capital)
Costs (annual costs to support the user in a given period)
SAC (subscriber acquisition costs, sometimes referred to as CAC = customer acquisition costs)
By using telemetry and modern data science each of the five variables can be optimized. Every process at the business, including but not limited to the sales funnel, can be tested and retested. Bill Gurley describes my view the complexity of this process far better than I ever have:
“Tren Griffin refers to the five variables of the LTV formula as the five horsemen. What he envisions is that a rope connects them all, and they are all facing different directions. When one horse pulls one way, it makes it more difficult for the other horse to go his direction. Tren’s view is that the variables of the LTV formula are interdependent not independent, and are an overly simplified abstraction of reality. If you try to raise ARPU (price) you will naturally increase churn. If you try to grow faster by spending more on marketing, your CAC will rise (assuming a finite amount of opportunities to buy customers, which is true). Churn may rise also, as a more aggressive program will likely capture customers of a lower quality. As another example, if you beef up customer service to improve churn, you directly impact future costs, and therefore deteriorate the potential cash flow contribution.
So what about MoviePass? Before digging into that question, I should say that I do not “pick stocks” on this blog. Revenue from this blog is < $0. No subscription fee is charged and there is no advertising. I am trying to teach people how to do their own valuations. I do have my own valuations, but I do not share them with readers. For example, when I wrote about a meal delivery business or a company like StitchFix I talk about a valuation methodology and approach that I think readers should consider using. But I do not assign the business a specific valuation. The approach that I describe involves a lot of research and detective work. To understand the stock, you must understand the actual business. If you do not like my approach of leaving the specifics up to you, you can have your $0 back.
To understand these model better I suggest you get a spreadsheet and play with the model yourself. Modify the assumptions and watch how it changes with each alteration of a variable. This is called a sensitivity analysis. This spreadsheet from David Skok of Matrix Partners is a great place to start your work. No, I won’t give you my spreadsheet or the simple one I used to create illustration in this post. No, I won’t do the work for you. as a side bar: Do you like to do this? Is it fun? Would you rather be doing something else? If you are not going to do this work yourself, you should buy a low cost and diversified portfolio of index funds instead of buying individual stocks.
Let’s assume MoviePass eventually has ongoing average revenue per user (ARPU) per month of $10.50. You can take that ARPU higher or lower in your modeling if you want. Knock yourself out on modeling different assumptions, but try to give yourself a margin of safety in the event you make a mistake. What is the probability that MoviePass will be able to generate significant data revenue on top of its break even movie subscription business? Ted Farnsworth, the CEO of the company that controls MoviePass (Helios and Matheson Analytics Inc. (HMNY)) is quoted in The New York Times as saying: “When you apply computer science and machine learning to an industry that we believe has lacked significant innovation, useful patterns start to emerge.” That is a nice theory. Will it work in practice? The HMNY CEO claims:
“18% of users go to movies after prompting by the MoviePass app. Making money putting people in the theater is fine, but also think about the advertising side,” Farnsworth said. “We’re the only company out there that can tell companies exactly who and when people are going to the movies.”
Sending more people to the movies with advertising seems like a mixed blessing given that each visit to a movie theater per month more than once is at least $9 in incremental CAC/COGS for MoviePass.
Remember that movie theaters are already operating at very slim margins. There are no extra profits to give away as one theater operators has made clear:
“MoviePass paid AMC, according to our records, $11.88 for each and every ticket that it purchased for our mutual guest. That’s quite a gap, $9.95 a month versus $11.88 a visit. I must point out that’s very gracious of them and we appreciate their business, but I think it’s also important to make clear that despite claims they’ve made to the contrary, AMC has absolutely no intention, I repeat no intention, of sharing any – I repeat, any, of our admissions revenue or our concessions revenue with MoviePass.”
Overall industry revenue can go up if spending by moviegoers increases to $9.95 a month, but at what cost? What is the return on investment?
Let’s assume MoviePass can get COGs down to $9.00 a month. That seems like a stretch given tickets cost an average of $8.93. MoviePass’s audience may be mostly urban residents who often pay something like $16 for a tickets as is the case in Los Angeles. MoviePass has other costs associated with its service that will be part of COGS that wil reduce gross margins too, like the cost of the debit card transaction etc.
Even assuming MoviePass eventually gets a gross margins up to a number like 18% that is a really skinny gross margin. A business will have a very hard time just paying marketing and general and administrative (G&A) costs with that gross margin, let alone any R&D.
What will the CAC of MoviePass be at scale? Well, it will include any sales and marketing costs plus the losses in the first few months when people see more than one movie a month that is in the COGS assumption. MoviePass illustrates a point I have made before: what seems like cost of goods sold (COGS) may really be CAC. People can argue about whether it is COGs or CAC from an accounting standpoint or how it should be reported, but what matters for investors is what the expense actually represents in the real world. COGS are the direct costs attributable to the production of the goods sold by a company. CAC is the cost associated in convincing a customer to buy a product/service. To illustrate, let’s assume MoviePass customers watch an average of 5 movies a month first month before settling down to consumption of one movie a month. $9 X 5 (the cost of the 5 extra movies in month one) is an addition to CAC of $45. What if it is 12 extra movies and not 5? 12 X $9 is a $108 addition to CAC. What if it is 20? Will that work financially? This sort of anecdote which appeared in a New York Times article about MoviePass is not comforting regarding CAC:
An anecdote is not data. If I could know one unknown fact about MoviePass it would be the level of fully loaded CAC. In the model below I assume CAC is $60 and then in another version $100.
Any lifetime value model is also sensitive to the amount of customer churn the service experiences. Customer retention is critical in a subscription business since churn can result in “stranded CAC” (i.e., the customer churns off the service before the operator gets back the cost to acquire that customer). Retention can be a really important variable since the best way to grow is not to shrink your existing customer base that you paid a lot to acquire. An executive associated with MoviePass has said publicly that retention per month is 96%, so monthly churn in this example is assumed to be 4%. There are a lot of ways to define churn and we could get technical about that, but let’s not. The goal is to be approximately right rather than precisely wrong. As Charlie Munger has said: “Using precise numbers is foolish. Working with a range of possibilities is the better approach. The decision should be obvious.”
The unit economics for MoviePass might look like this below. All inputs in blue are educated guesses.
What if CAC $100? Things can get ugly fast (or not):
Does MoviePass really have product/market fit given the existing unit economics? Or are they trying to find a solution to a growth hypothesis before the value hypothesis has been proven and face all the dangers that can come with premature scaling? A key part of the value hypothesis is a sound business model and that includes sound unit economics. Sound unit economics are not always fully developed when growth investments start, but they need to be well along the way to being proven. “Build it and the unit economics will come” is a dangerous game.
A business that is seeking growth for its own sake is a bad idea. No one has made this point better than Warren Buffett: “Growth is always a component in the calculation of value, constituting a variable whose importance can range from negligible to enormous and whose impact can be negative as well as positive.” If the startup does not have have product/market fit and yet is trying to grow they are telling customers that the product is not really valuable. Revenue alone is not enough to establish that product/market fit exists/ a value hypothesis has been proven sufficiently. What would happen if customers are asked to pay a higher price point for the same subscription is always unclear. What is known is that consumers hate to lose something they already believe they have. John Malone hates to lower prices since raising them after that price drop is hard.
Like any subscription business MoviePass must be laser focused on keeping CAC and churn at attractive levels. In the case of a business like MoviePass which has low gross margins, high CAC makes high churn a particularly dangerous phenomenon. My educated guesses about CAC and churn at MoviePass in the spreadsheets are just that, educated guesses. It is worth noting that when a business publicly or privately reveals an input into lifetime value like ARPU, CAC or churn it is almost always an average. That may hide the fact that there are big differences in the analysis by “cohort.” A cohort is a collection of customers who share an attribute or set of attributes. For example, one type of a cohort would be those customers who subscribed to a service in a given quarter or year. Early MoviePass customers may be quite different than customers who arrive later. A separate a LTV analysis for each cohort should be created.
If a business pays enough CAC, it can almost always generate a lot of revenue growth. Ice can be sold to Eskimos if you make the sales commissions and customer incentives high enough. MoviePass recently announced that it has passed the 1 million mark in paid subscribers. Did they do this by selling a dollar for significantly less than a dollar? Or it’s a workable business model? That determination can only be made by looking at the LTV model and you can only do that if you have data I do not have. How many movies do members see? Do they settle down to an average of one per month? If so, when? Without knowing CAC, customer churn and what the cash flow statement looks like, it is really hard to value the MoviePass business. You can do some rough math even without all the data by making some assumptions based on comparable businesses, but you are guessing. The more businesses like this you have seen in a lifetime in more industries, the better you get at making that educated guess.
The cash flow aspects of the upfront CAC are important to understand. For example, monthly positive cash flow in my first example screenshot of the spreadsheet above is only $1.89 given my assumptions, but $60 in cash has spent on CAC already. This creates a significant a hit to cash on hand. The faster you grow the subscription business, the faster cash flies out the door. A subscription business like I describe above almost always eats cash and requires outside money to deal with the negative cash flow if the business is growing. Eventually a subscription business can flow more cash than it consumes, but that can take a very long time. An important question always is: how long does that journey to positive cash flow take? There are two important rules in this situation: (1) watch cash carefully and (2) watch cash carefully. If the business runs out of cash, it has committed the only unforgivable sin in business. Don’t do that.
When mobile and cable TV operators were building their businesses Michael Milken was there supplying the growth capital. Who is going to supply the growth capital (cash) to MoviePass? MoviePass sold a majority stake to the listed company HMNY in August for less than $30 million. This is not a traditional approach to acquiring growth capital and that is not a lot of money in the world of subscriptions. High growth subscriptions businesses require a lot of cash for a long time. MoviePass selecting HMNY was unusual? The company does not have the deep pockets of Fidelity, Silver Lake, TPG or Blackstone in terms of resources. How much cash will MoviePass need to get to cash flow break even? HMNY recently raised $60 million for the expansion of MoviePass, but in the world of subscription $60 million disappears like water in the desert. Do you want another metaphor? OK. Up front CAC eats cash like the Cookie Monster eats cookies. Bill Gurley writes: “Let’s say you have a company that estimates it will do $100mm in revenue this year, $200mm the next, and $400mm the year after that. In order to accomplish those goals it is going to invest heavily in marketing – say 50% of revenues. So the budget for the next three years is $50mm, $100mm, and $200mm.” That’s a lot of cash.
When thinking about a business like MoviePass, it is a good idea to keep in mind what Zhou Enlai is often quoted as saying when asked his opinion of the French Revolution’s effect on world history. “Too soon to tell.” Yes, he probably was not referring to the French revolution, but the apocryphal remark is, as the FT once wrote, “too delicious to invite correction.” Unless you are a member of the MoviePass management team, a consultant who had signed a NDA, the auditor or an investor in the company, no one knows enough about MoviePass’s operating metrics and financials to make a firm conclusion. But there are clues which allow you to make different models as I have above. I have hopefully given you some new tools and ways of thinking that can help you make a more informed analysis.
P.s. I wrote about the rise of subscription business models in my blog post “Amazon Prime and other Subscription Businesses: How do you Value a Subscriber?” https://25iq.com/2017/07/15/amazon-prime-and-other-subscription-businesses-how-do-you-value-a-subscriber/ and “A Half-Dozen Ways to Look at the Unit Economics of a Business.” https://25iq.com/2016/12/31/a-half-dozen-ways-to-look-at-the-unit-economics-of-a-business/
Slate on movie economics: http://www.slate.com/articles/arts/the_hollywood_economist/2006/01/the_popcorn_palace_economy.html
Wall Street Journal https://www.wsj.com/articles/SB94961357469579449