Netflix Elicits More Anger, Fewer Google Searches

Nov. 2.11 | About: Netflix, Inc. (NFLX)

It took me $7.64 of red spray paint, an old moving box, some white paper and a patient and crafty fiancée to find out how uncool Netflix really is. For Halloween this year, my wife-to-be decided that I would go as Netflix, since I'd spent the last month talking about them. Using the aforementioned supplies, she diligently made me a perfect replica of a Netflix DVD sleeve that I wore for a Halloween party and then to a bar.

I was Netflix. Never have I felt so uncool. To be honest, there were times I was even a bit scared. Ten people came up to me and told me how they were canceling their plans, three shouted "uncool!" at me, one creepily told me he knew where I live (I am sorry to the person whose address I used on the sleeve!), and there were a few that said things unfit to print. I had never met any of these people. There were some compliments on the costume design, but these were each tempered with negative remarks about the company itself.

People at a bar on a Monday Halloween in Chapel Hill, N.C. are hardly representative of the target Netflix customer. Nonetheless, I do not think Netflix is in a good position going forward. In the Q&A after the recent Q3 conference call, CEO Reed Hastings stated: "Now, what we've seen is a second wave of cancellations from the price increase. The first wave was in July upon announcement, and the second wave has been in September and October, as people become more aware of the price increase, and then either changed the plan or canceled. And that wave has been declining very steadily over the past couple of weeks." Given the reaction to my costume tonight, I think he might want to worry about a third wave.

In my last article on Netflix, I talked about two worrying trends for Netflix. The first was a substantial decline in numbers of free subscribers in Q3. This is especially problematic because free subscribers in the current quarter have, historically, been an excellent guide to paid subscribers in the next quarter. In the conference call Q&A, Hastings acknowledged this role: "Our marketing has been very successful for the last several years, and we don't plan on any substantial change to it. It's a great set of campaigns that work very well at attracting streaming subscribers. And in Q4, particularly in December, there's more focus on a set of devices, video game consoles that get sold, iPads that get sold, a wide range of devices. And so our expectations are modeled from [the] prior year's performance."

What happens if, as investors, we model expectations from the Q3 marketing campaign - which I am going to measure as free subscribers in the current quarter - from last year's performance? Columns 1 and 2 of Table 1 give the number of paid and free subscribers by quarter in 2010 and 2011. They show pretty clearly how Netflix ramped up its marketing in 2010 and how successful it was in attracting new paying subscribers in each quarter.

Column 3 is what makes Table 1 interesting. It shows the quarter-on-quarter increase in paying subscribers for each free subscriber Netflix had at the end of the prior quarter. I call this marketing conversion, because free Netflix has historically focused on offering free one-month trials. (Lately, I have also heard many commercials on the radio).

I draw readers' attention to column 3 because I think it shows a relatively consistent level of performance over the last year. Of course, the most recent quarters have been affected by losses, as the quote from Hastings details, but generally Netflix achieved between 2 and 3 converts for each free subscriber over 2010. I assume that this is what he means when he says "our expectations are modeled from [the] prior year's performance.".

Table 1: End-of-quarter paid and free subscribers and marketing conversion, 2010 to present




Thousands Paid

Thousands Free

Marketing conversion





























Click to enlarge

Source: Netflix 10Qs. Marketing conversion defined as Q-on-Q difference in paid subscribers divided by prior-quarter free subscribers. For example, 2.77 in Q2 2010 means that Neflix signed up 2.77 people for each free subscriber in Q1 2010.

Whatever multiplier one wants to use, this does not bode well for Netflix's potential Q4 subscriber numbers, , because the number of free subscribers in the current quarter dropped so substantially.

The second worrying trend for Netflix was changes in search patterns for the company. In my last article, I showed an extremely strong association between Netflix's subscriber numbers and the density of Google searches for the term "Netflix," using Google Trends. I assume that many people search for the company's website rather than enter it directly into their browser's address bar when they want to update their queues or watch something instantly.

Google Trends data are updated weekly (and slight revisions appear to be made to prior weeks), so I am dedicating the remainder of this article to updating my prediction of the number of unique domestic paying subscribers for Netflix in Q4. Because the subscriber data come in weekly segments, I use a simple average of the number of searches in a quarter. I then use linear regression to associate the number of paying domestic subscribers in historical quarters with searches for "Netflix." I control for searches for "NFLX" and "Netflix price" to account for rising interest in the stock over time and news articles about the company's recent price increases, which may be driving changes in who is searching for Netflix and thereby bias the results.

Figure 1 updates the subscriber predictions from my prior article using the revised data and incorporating one more week of Q4 results.

Figure 1: Predicted vs. actual domestic paying subscribers, based on Google Trends

Click to enlarge

The additional week of data did not change my sentiment. Indeed, whereas I had previously predicted 20,734,500 subscribers based on Google search trends, the new week of data and revisions to prior data suggests that 20,637,310 is the best guess for Q4.

I am well aware that this is a simple analysis, and more rigorous methods could be used, but it has had so much predictive validity in prior quarters that I believe more people should be aware of it. The decline in Netflix's popularity among Google searches continues and, as my recent Halloween experience shows, there is a lot of anger out there.

As a final thought and reminder, I wanted to highlight how important paying domestic subscriber numbers have been for the value of NFLX shares in the past. Figure 2 shows how changes in Netflix's stock price over the next quarter associate with changes in paying domestic subscribers in the current. Given figure 2, I would not go long NFLX until I see it increase subscriber numbers substantially. Others have given more nuanced analyses of various problems in its business model (I, II, III, IV), but I think Netflix needs to drastically increase its marketing and its effectiveness to ever grow value for shareholders again.

Figure 2: Paying domestic subscriber numbers and next-quarter change in stock price

Click to enlarge

Disclosure: I am short NFLX.