The other day I let my 7 year old daughter use my notebook, something that rarely happens. While playing a game where she dresses up unicorns, I noticed Google (GOOG) was serving ads in the sidebar for "Incredible Stock Market Secrets". The ads were obviously aimed at me, not a child engrossed in unicorn fashion.
That got me thinking about Google's data mining ingenuity, not unicorns. Any other company probably would have served an ad related to the children's game site. Ads served to children might get occasional click throughs, but deliver a very low return on investment. Although my 4 year old accidentally bought $10 worth of MobaCoins from the Play Store last year, most children have very limited purchasing power. Thanks to Google's algorithms, both the games for children website owner, and the dodgy stock market tips advertiser benefited from the algorithm that aimed an ad at my interests, rather than the site's content.
Over the past year, I've noticed a rising interest in the traffic acquisition costs, referred to as TAC, of various internet companies and how they relate to overall profitability. What I haven't seen is a side by side comparison, so I spent some time with the latest 10-K filings of the largest internet companies I could think of that rely on advertising for the majority of their revenue.
A Brief Explanation of TAC
For internet companies like Baidu (BIDU), Facebook (FB), Google, and Yahoo (YHOO), advertising comprises the vast majority, if not all, of their revenues. TAC is the percentage of those advertising revenues paid to third parties that direct surfers to their websites or advertisements.
As you read this post, there are probably two or three ads from Google somewhere on the page. If you click one of them, Google will charge the advertiser a certain amount, let's say one dollar. Google then credits Seeking Alpha with an undisclosed percentage of that dollar. If Google credits the Seeking Alpha account with $0.50, the TAC is 50%.
It makes sense that TAC should be an important indicator of profitability for internet companies that rely on advertising for the majority of their revenues. I'm going to show you that is not necessarily true.
First, let's take a look at advertising as a percentage of overall revenue for 2012.
Source: company SEC filings
Percent of Total Revenues
As you can see Baidu relies almost entirely on advertising to generate revenues. Facebook is steadily increasing the amount of revenue it generates from payments to its platform developers like Zynga (ZNGA), but still relies heavily on advertising. Google's physical products might be getting the most attention from news outlets, but advertising still dominates its revenues. Finally, Yahoo's tangled mess of advertising still comprises just over 80% of the company's total revenues.
Microsoft (MSFT) is not included in the above table because online advertising is such a small part of its total revenues. During Softy's fiscal year 2012, which ends June 30, online advertising generated just $2.6 billion. That's only 3.5% of the $73.7 billion the company generated in revenues.
Let's Take a Look at TAC for these companies during fiscal 2012.
Source: company SEC filings and YCharts
TAC As Percent of Ad Revenue
Overall Operating Margin TTM
Baidu relies heavily on its Baidu Union members for revenue, but apparently doesn't pay them very much. Similar to Google Adsense, Baidu Union members operate their own websites and mobile apps that direct traffic to Baidu. Although the company stresses the significance of its affiliates' contribution, I couldn't find any information regarding the percentage of ad revenue originating from Baidu's own properties versus those of Baidu Union members. The Chinese company's latest 20-F filing is well over one hundred pages, so I may have missed it. If you have the data please comment.
Google's Adsense program has a user interface available in dozens of languages, including Chinese. The company also breaks down the proportion of ad revenues generated on its own websites and those of affiliates. For the year ended 2012, $12.465 billion, about 27%, originated from "Google Network Members' websites." Sadly the company doesn't break down ad revenue figures by country in its SEC filings. Again, if you know how much of Google's ad revenue is coming from China, please comment.
Yahoo's shotgun approach to generating ad revenue is frustrating to say the least. Just try to wrap your head around this quote from the company's 2012 10-K:
Our agreements with Microsoft and AOL (AOL) allow ad networks operated by Yahoo!, Microsoft and AOL to offer each other's premium non-guaranteed online display inventory to their respective advertising customers.
I can only assume that the $519 million the company spent on TAC last year was paid to both Microsoft and AOL.
Yahoo had an affiliate program for individuals and small businesses similar to Adsense, but as far as I can tell, it isn't operational any longer. Years ago, Yahoo partnered with Commission Junction, owned by ValueClick (VCLK), to provide an affiliate program for its web hosting, and other services. The affiliate program's web page is still live, but hasn't been updated since 2009. Hopefully, for Yahoo investors, Marissa Mayer can bring to its affiliate program the streamlined simplicity that both small and large companies enjoy from Google Adsense.
Facebook doesn't even mention traffic acquisition costs for it's ad revenue, because it essentially has none. All of its ads appear on its own site. Although I have seen some ads inviting surfers to join the social network, those expenditures would no doubt be lumped together under marketing and sales.
You're probably wondering why I included Facebook in this post despite its conspicuous lack of TAC. It serves as a perfect example of how this often overhyped metric doesn't necessarily relate to an internet company's profitability. Last year the social network behemoth cleared just $53 million in net profits from $5.089 billion in revenues.
I should note here that although Facebook has no TAC, its overall operating margin is surprisingly slim. This is largely due to its aggressive R&D program consuming a whopping 27% of total revenues. That's almost $1.4 billion going to some very excited engineers working on a single internet property.
Google's TAC related to its affiliates has been dropping, but distribution related TAC is on the rise. In 2012, Google paid Apple (AAPL) approximately $3.2 for every iPhone and iPad sold, in return for being its default search engine. That might sound like a lot, but its really only several clicks per device. It also paid Mozilla, makers of the Firefox Browser, an estimated $200 million last year to be its default search engine. During the latest earnings call, CFO Patrick Pichette was wonderfully coy when explaining the rise in distribution related TAC. Without naming Apple directly, he said "the volume of mobile in our mix is driving TAC to go up."
Patrick's argument for allowing distribution TAC to rise is one that I fully agree with. Even if paying a great deal to have a presence on every possible device isn't directly profitable, it creates a positive feedback loop. If Google outbids competitors to create a ubiquitous presence, it has more user data to work with. With more data, it can deliver a better experience for users. By delivering a better experience, it generates more revenues. With increased revenues, it can continue to outbid competitors and maintain a ubiquitous presence.
Keeping an eye on TAC isn't the worst thing an investor can do with their time, but don't miss the forest for the trees. A rise in TAC among smaller internet companies in a price war is definitely negative. However, allowing TAC to rise in order to expand your economic moat is most likely a good thing.