Are Netflix's Lost Subscribers Going To Redbox? (Part 1 Of 3)

Includes: NFLX, OUTR
by: Michael Verd

In part 1 of this article, I set up my reasons for investigating Coinstar (NASDAQ:CSTR) stock and set up what I see to be the key question that will determine Coinstar returns in the coming quarters: Whether Netflix's (NASDAQ:NFLX) lost customers are opting for Redbox. In the next installments, I look at what state trends in Google searches might have to say about this question and try to put it all together and reflect on how this analysis has informed my position with respect to Coinstar stock.


I usually try to make investment decisions after careful research. However, sometimes I act on my gut and what I assume to be a logical deduction. I did this last week when I bought some shares of Coinstar before its earnings announcement. My gut told me that kiosks are a profitable model, one that can easily be expanded in interesting new directions. I pulled up Coinstar's 10Qs and noted a considerable growth in the number of Redbox kiosks installed of late, and I convinced myself with the logical argument that, given Netflix's recent loss of subscribers and general anger toward the company, Coinstar was in a nice position to benefit.

These are the types of investment decisions I try to avoid - those that on face value seem reasonable but which in reality are not based on anything more than a hunch - but I still make them sometimes. I did not lose too much money with this one, but I still hate violating my principles.

With that in mind, I decided to further interrogate what I see to be the key question for the near term performance of CSTR stock: Are Netflix's lost subscribers opting for Redbox? Granted, there is a lot to like about the company in general - the green coin machines are a brilliant invention - and they seem poised to move in some exciting new directions. But, in the near to mid-term, whether they can capitalize on Netflix's recent hiccups seems likely to drive the market's valuation of them.

The market seems to be betting that they cannot. With a P/E of 14.39, they're right around the market average. With a days to cover ratio of 8.9, there seems to be a lot of short sellers betting against them relative to the number of shares traded on a typical day. (Days to cover is measured as the number of share short over the average daily trading volume of the stock). For comparison, here are the days to cover ratios of some high-flying stocks that many people are shorting: NFLX (.8) , AMZN (1.7), MSG (3.0), CMG (3.2), GMCR (3.6), CRM (4.6), NYT (6.2). (I believe one typically takes days to cover as a contrarian indicator since short sellers taking profits on their position adds a floor to the stock price falling and the buying demand from them covering their losses adds jet fuel if it rises ... please explain to me in the comments if I am misinterpreting this).

General Approach

The remainder of this article will try to answer whether Netflix's lost subscribers are going to Redbox. This article is about Coinstar's position in the circumstance created by Netflix, and, as such, I do not intend to make any comment about what the results of my analysis imply for Netflix's future. For one thing, the dynamic between Netflix and Redbox is not zero sum. At least, many people I know seem to use the two services together, getting newer releases on a whim from Redbox and older releases more deliberately from Netflix. My guess is that the recent abandonment of Netflix by many subscribers will result in more people trying Redbox and, for most of them, sometimes coming back. To be honest, this is what I hope to find, confirmation that my gut was right and that I made the right decision.

My last articles made the case that Google trends data can sometime be useful in understanding a company's growth prior to earnings releases. Of course, this is only going to be true for companies/products that people might search the web for. I doubt there is much benefit to incorporating search trends for military equipment or cancer drugs in trial stages, but consumer products and services will likely be reflected in search trends.

In this analysis, I make the assumption that peoples' online search behavior for the terms "Netflix" and "Redbox" bears some relation to their use of the products. I think it is surprisingly common to Google a website one wants to visit rather than typing it in directly to the address bar. For Netflix, people might be searching the website so they can use it to stream videos or update their queue. For Redbox, people might be looking for nearby locations or checking to see what titles are available nearby. (If you do not buy this assumption, you might want to skip to part three of the article then come back.)

Of course, more people must interact with the Netflix website to use their services (streaming or DVD ordering), while many can make a decision in the grocery store line about Redbox. And many access Netflix through third party devices. These factors make this an imperfect method of detecting trends in customer behavior around these two companies. Nonetheless, I still expect that some signal will be visible through the noise.

To try and find some clarity about what lost Netflix subscribers might be doing with respect to Redbox, I downloaded the search trends from 2004-present for the terms "Netflix," "Redbox" and "IMDB" separately by each state (and D.C.). I hope that the first two terms will capture some dynamic of what is occurring with the two companies since Netflix outraged customers with its price increase, while I intend the last to be a control for temporal and spatial variation in peoples' interest in movies generally. What is nice about this approach is that I can exploit the space-time variation in trends to try and answer my question. (I discuss why this is beneficial in greater depth in part 2 of the article.)

Comparing Five States

My first pass at analyzing whether Netflix customers are opting for Redbox is purely descriptive. I compare search patterns for "Netflix," "Redbox" and "IMDB" in five states shown in figures 1-5. I picked these states for a variety of reasons: They are all in the top 10 most populous, their search data is reasonably complete, they are in several areas of the country, and the trend lines in their search terms show some heterogeneity in recent dynamics.

Trends in California, the country's most populous state, are shown in figure 1. For most of 2011, search levels for Netflix and Redbox tracked each other closely. However, right around the beginning of Q3 2011, there is a marked divergence. From that point, Netflix searches began to fall while Redbox searches continued to trend slightly upward. Because of how closely this aligns with Netflix's announcement of price increases, it certainly suggests that Netflix customers might have been leaving for Redbox in California. There is no notable change in search levels for IMDB at this point in time, which seems to supplement the argument because - were something else going on (e.g., lack of interest in movies in the summer) - we might expect it to show up there.

Figure 1. Recent search trends in California, population in 2010: 37.3 million.

Click to enlarge

Figure 2 shows trends in New York, which tell a different story. The slowing in Netflix searches in New York begins at almost the same time as it did in California, but the decline is much less pronounced. However, unlike in California, Redbox searches show no appreciable uptick after Netflix announced price increases. If anything, there appears to have been a decline, albeit one that is parallel to the slight decline in IMDB searches over the summer and early fall.

Figure 2. Recent search trends in New York, population in 2010: 19.4 million.

Click to enlarge

In Florida, shown in figure 3, we have yet a different trend. Again, Netflix searches took a large hit in the most recent quarters. Here again, however, Redbox does not appear to have been a substantial beneficiary. While Q3&Q4 of 2011 show a slight increase in search volume for Redbox, it is less than I would expect given the declines in Netflix searches. Interestingly, there is a run up in searches for both Netflix and Redbox in Q2, but both trends appeared to hit a wall in Q3 (though Netflix hit a bigger wall).

Figure 3. Recent search trends in Florida population in 2010: 18.8 million.

Click to enlarge

Figure 4 shows Illinois whose search patterns seem similar to Florida, except that the decline for Netflix and the run up in both terms in Q2 are more attenuated. In Illinois, there is a large jump in searching for Redbox just before Q4, which may be a result of people canceling their Netflix accounts when the price increases began in September, but it is too early to tell if this will be a sustained trend.

Figure 4. Recent search trends in Illinois, population in 2010: 12.8 million.

Click to enlarge

Finally, figure 5 portrays yet another pattern in Georgia than those seen in other states. In Georgia, Netflix searches have crashed substantially since the price increase, while Redbox searches have risen substantially. However, the rise in Redbox prices in GA appears to be more of a continuation of a prior trend than anything notably different than what was seen before Netflix announced it was raising its prices.

Figure 5. Recent search trends in Georgia population in 2010: 9.7 million.

Click to enlarge

These five states illustrate a potentially surprising heterogeneity in trends since Netflix announced price increases in July. Only one thing is constant across graphs: Netflix searches declined substantially in every state almost immediately after they announced price increases. I have already spelled out why I think this is bad news for Netflix. However, the stated goal of this article is to evaluate whether search trends can reveal anything about whether Netflix's lost subscribers are headed to Redbox. Based on the results of this state level comparison, I see mixed evidence for this occurring.

In the next section, I try to incorporate information from more states to see if considering the entire country together gives any clue.

>> Continue to Part II

Disclosure: I am short NFLX, CSTR.

Additional disclosure: I am short both through puts.