Digging Into Netflix's Marketing Expenses

| About: Netflix, Inc. (NFLX)

I previously wrote an article looking at Netflix, Inc.'s (NASDAQ:NFLX) recently released financial statements, with a focus on the company's international segment. NFLX's 10-Q became available Wednesday at www.sec.gov. The 10-Q contains a lot of useful information, but I wanted to examine the company's marketing expense and in particular the implications of its 1-month free trial subscriptions. According to the 10-Q:

Marketing expenses consist primarily of advertising expenses and also include payments made to our affiliates and consumer electronics partners and payroll related expenses. Advertising expenses include promotional activities such as television and online advertising as well as allocated costs of revenues relating to free trial periods. - NFLX Q1 2011 10-Q

The first question here is to dig a little deeper into the number of free subscribers. The following table shows a calculation of the number of free subscribers over several of the previous quarters:

NFLX Free Subscribers (Thousands)
Period Beginning Period Free Ending Period Free Estimated Average
Q1 2011 1,742 1,572 1,657
Q4 2010 1,070 1,742 1,406
Q3 2010 424 1,070 747
Q2 2010 345 424 385
Q1 2010 376 345 361
Q4 2009 274 376 325
Q3 2009 224 274 249
Q2 2009 194 224 209
Q1 2009 226 194 210
Source: NFLX SEC filings. The estimated average is just the straight average of the beginning period and ending period free subscribers.

The first observation is that the number of free subscribers increased significantly by the end of Q3 2010. NFLX attributed this to a significant increase in the use of 1-month free trials instead of the more traditional 2-week free trials. The following note (in some form) appeared in the Q1 2011 10-Q and the 2010 10-K but not the Q3 2009 10-Q

The 3.9 percentage point increase in free subscribers as a percentage of total subscribers as of March 31, 2011 as compared to March 31, 2010 is due to the expanded use of our one month free trial subscriptions over the previously used two week free trials. - NFLX Q1 2011 10-Q

When I first joined NFLX, I had a 2-week free trial. In searching the internet, I found references to 1-month free trials as early as Q2 2008. However, it appears they were not very common until just very recently and perhaps have become the only free trial period today.

However, a key aspect of NFLX's performance is its ability to continue its strong growth rate. It seems most analyses of NFLX focus on growth at either a quarterly level or even an annual level. It looks at how many millions of net subscribers join over some time frame.

However, digging into the free subscriber numbers might reveal some thoughts on daily join rates. First, though, I'll start with an assumption that free subscribers are ultimately pretty regularly converted into paying subscribers. Hence, a decline in the acquisition rate of these subscribers will eventually impact the growth of paying subscribers.

Unfortunately, NFLX does not publish this data in any form. The company simply provides point estimates for a few days each year. However, with some basic assumptions, I can dig a little deeper. On March 31, 2011, NFLX had 1,572,000 free subscribers. Depending whether these subscribers had a 2-week free membership or 1-month free membership, this means that they joined within either 14 days or 31 days. Assuming that everyone had 31 day free trials and people joined uniformly each day this would mean that every day 1,572,000/31 = 50,710 people joined. So the first question is how does this compare to previous quarters:

Daily Acquisition of Free Subscribers - Base Case
Period Free Subscribers (000) % 2 weeks % 1 Month Blended Free Period (days) Implied Daily Add Rate (000)
Q1 2011 1572 5% 95% 30.2 52.1
Q4 2010 1742 15% 85% 28.5 61.2
Q3 2010 1070 25% 75% 26.8 40.0
Q2 2010 424 95% 5% 14.9 28.6
Q1 2010 345 95% 5% 14.9 23.2
Q4 2009 376 95% 5% 14.9 25.3
Q3 2009 274 95% 5% 14.9 18.5
Q2 2009 224 95% 5% 14.9 15.1
Q1 2009 194 95% 5% 14.9 13.1
Q4 2008 226 95% 5% 14.9 15.2
Source: Derived from data in NFLX 10-Q and 10-K filings. My estimates for percentage of 2-week and 1-month free trials. The clear caution is that I'm really guessing at the percentages. The math in this table works with 1,572/30.2 = 52.1

The first observation over the longer term is that NFLX shows strong growth in acquisition of free subscribers. However, it appears that the growth has declined significantly from Q4 2010 to Q1 2011. While the number of free subscribers declined overall, if there were even more 1-month free trial subscribers than in the previous period, this daily decline becomes even more pronounced. The following two tables show some other percentage examples to illustrate this.

Free Subscriber Acquisition Rate - Case 1
Period Ending Free Subscribers (000) % 2 weeks % 1 Month Blended Free Period (days) Implied Daily Add Rate (000)
Q1 2011 1572 0% 100% 31.0 50.7
Q4 2010 1742 25% 75% 26.8 65.1
Q3 2010 1070 50% 50% 22.5 47.6
Source: same as previous table.

This first case shows an even more dramatic decline, showing free subscriber acquisition rates falling to close to the Q3 2010 level. The next case will be with constant percentages across the last three quarters, which will also show a decline in the daily acquisition rate.

Free Subscriber Acquisition Rate - Case 2
Period Ending Free Subscribers (000) % 2 weeks % 1 Month Blended Free Period (days) Implied Daily Add Rate (000)
Q1 2011 1572 10% 90% 29.3 53.7
Q4 2010 1742 10% 90% 29.3 59.5
Q3 2010 1070 10% 90% 29.3 36.5
Source: same as previous table.

So it appears that the decline in acquisition of free trial customers could be somewhat significant.

Drastic Increase in Spend on Free Trial Subscribers?

The next goal of this analysis is to break out the marketing expense into spend on serving free trial customers-- as opposed to other marketing expenses-- which would include payroll and traditional advertising. The first step is to determine the cost of serving a typical customer, with the assumption that it costs the same to serve a paying subscriber as a non-paying subscriber. The following table breaks out cost of revenue per day per paying customer:

Cost to Serve Average Subscriber
Period Cost of Revenue ($000) Average Subscribers (000) Cost to Serve per Quarter ($) Cost of Revenue per Day ($)
Q1 2011 438,151 19,670 22.28 0.244
Q4 2010 390,790 18,472 21.16 0.232
Q3 2010 344,469 15,220 22.63 0.248
Q2 2010 314,934 14,100 22.34 0.245
Q1 2010 307,162 12,757 24.08 0.264
Q4 2009 275,486 11,689 23.57 0.258
Q3 2009 275,274 10,605 25.96 0.284
Q2 2009 269,243 10,246 26.28 0.288
Q1 2009 259,268 8,169 31.74 0.348
Source: NFLX 10-Qs, average subscribers for Q4 2010 and Q4 2009 taken as straight average of period end and period beginning paying subscribers. Assumes 91 days in a quarter.

While the goal is to find the cost of revenue, there has historically been a downward trend. However, these costs did increase into Q1 2011 (but also increased into Q1 2010). The next table will look at marketing expense and breakout an estimated cost of serving the free trial subscribers.

Marketing Expense for Free Subscribers
Period Marketing Expense ($000) Average Free Subs (000) Cost of Revenue per Day ($) Cost of Revenue for Free Subs ($000) Other Marketing Expense ($000)
Q1 2011 104,259 1,657 0.244 36,809 67,450
Q4 2010 62,849 1,406 0.232 29,664 33,185
Q3 2010 81,238 747 0.248 16,860 64,378
Q2 2010 74,533 385 0.245 8,565 65,968
Q1 2010 75,219 361 0.264 8,656 66,563
Q4 2009 70,715 325 0.258 7,639 63,076
Q3 2009 58,556 249 0.284 6,446 52,110
Q2 2009 46,231 209 0.288 5,477 40,754
Q1 2009 62,242 210 0.348 6,647 55,595
Source: Marketing expense is from NFLX Q1 2011 financials on their website. Average free subscribers is based on one of my earlier calculations. Cost of revenue per day is based on the previous table. Total Cost is just 91 x Cost per Day x Average Free Subs. 91 is the estimated days per quarter.

So it seems like an enormous amount of the marketing budget is going to support free trial customers, substantially more than in previous quarters and growing at an alarming rate. In contrast, all the other marketing expenses remain somewhat constant, with a strange dip in Q4 2010. While this is interesting, one part of the 10-Q that was puzzling to me was the explanation of the increase in marketing expense from Q1 2010 to this quarter, as quoted below:

The $29.0 million increase in marketing expenses was primarily attributable to a $25.1 million increase in marketing program spending, primarily from increased spending in television and radio advertising, coupled with an increase in our online advertising and payments to our affiliates. These increases were offset by a decrease in partner programs and inserts. Domestic subscriber acquisition cost decreased primarily due to continued strong organic subscriber growth. - NFLX Q1 2011 10-Q

I don't find this statement consistent with my analysis, which suggests that I'm not interpreting it correctly. Perhaps marketing program spend covers the costs of free trials - or perhaps there is some flaw in my analysis. However, it seems like there should be a substantial increase in cost of service to free subscribers and my methodology seems reasonable. One clear issue is that the number of free subscribers could vary significantly within a quarter.


This deeper dive into NFLX's marketing expense seems to suggest that sustaining customer acquisition is becoming both harder and quite possibly more expensive. NFLX seems to be spending much more on free subscribers than ever before. Also, overall acquisition costs for subscribers showed a large increase from Q4 2010 to Q1 2011.

Disclosure: I have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

Disclaimer: This article is for informational and educational purposes only and shall not be construed to constitute investment advice. Nothing contained herein shall constitute a solicitation, recommendation or endorsement to buy or sell any security.