CandidCapital

As Google's grip on the on-line advertising market gets ever tighter - who is set to be the real beneficiary of the long tail?

One of the ideas to come out of Xerox Corporation's Palo Alto Research Center (PARC), source of so much that has shaped modern computing - from the mouse to object oriented programming, is 'pervasive computing' - the idea that everything you do is impacted by and interacts with computers.

In the internet it seems that Google has achieved that status. 'Googling' has entered the dictionary, reflecting the fact that most people go to Google to locate information on the internet. Sure, if you know what you're looking for, you can go straight to the right place: BBC News, Yahoo! Finance, Wikipedia or even CandidCapital. But where the information is disparate or located in many different places, Google is the service that most of us will turn to for help.

At the risk of stating the obvious, Google takes the search term (the words) entered by a user and returns web pages (or blogs or images increasingly) which have key words matching that term (or its constituent words). On presenting the results of the search, Google tells you how many web pages have key words that match the search term and lists hyperlinks to the first ten of those pages (unless you tell it otherwise). Two points to note: first, that Google's 'secret sauce' - the PageRank algorithm - has determined which pages will be in the first ten results and the order in which all the other results are listed and, secondly, that the ten listed links aren't all that's presented on the page.

PageRank is a sophisticated algorithm orginally developed by Google's founders, Larry Page and Sergei Brin, when they were PhD students at Stanford University. It attempts to determine the 'relevance' of each page found containing key words that match the search term, using comparative relevance to determine a ranking that is used to display the results of the search. Herein lies the tension between search engines (since they all use some form of ranking algorithm) and publishers of web pages. Publishers want their pages to come high up in the ranking of pages returned when the user enters a search term that publishers see as relevant - irrespective of the user's view of relevance. Search engines, in contrast, want to deliver results that the consumer sees as relevant - or do they?

Clearly, if a web page is designed to sell something, there's a real financial advantage in getting a place high in the ranking determined by PageRank. Getting into the first ten results - displayed on the first page shown to the user - enormously improves the odds that the user will visit the publisher's site and make a purchase. By the time of the dot.com bubble in 1997-99, consultants were offering to tweak web pages to improve their rankings on all the leading search engines (which were Alta Vista and Lycos in those pre-Google days).

Google was a relatively late entrant to the search engine space, spinning out of Stanford in 1997/8. It grew rapidly, partly on the back of its clean user interface and the effectiveness of PageRank and partly through the partnership it had with Yahoo! (Google provided the search functionality to the Yahoo! site from 2000 until 2004). In the results for the last full year prior to its IPO (year to 31 December 2003) Google's sales were $962M and its net profit $106M. In its most recent quarter (to 30 Septmber 2007) sales were $4.2Bn and net profit $1.1Bn - equivalent to $16.9Bn sales and $4.28Bn net profit in a full year.

What is it that accounts for this extraordinary growth in sales (17x over 4 years) and profit (40x over the same period)? AdSense and AdWords, of course, which brings us neatly back to the second point of note about Google search results. To the right of the ten top ranking search results are up to eight "sponsored links" and above them up to a further two sponsored links. These are short text adverts that are relevant to the search terms entered by the user. Nor is it only on the Google site itself that they appear - millions, probably tens of millions, of sites carry AdSense adverts - each block headed with 'Ads by Google'. It is an impressive tool - the adverts are extremely well-focused: on our sister PC advice site, SortedPC.net, I blogged about the challenges of calendar sharing and the ads in the AdSense block that shows with that page are invariably highly relevant. Clearly this works especially well because of the massive online advertising activity of the computer industry. Had the blog been about, say, the Women's Institute, AdSense would have had to find other themes in the text to drive the advertising.

Google's Q3 2007 financial statement reveals that 34% of total revenue ($1.45Bn) came from AdSense on partner sites. That's nearly $6Bn a year - but what do the sites that carry the advertising get? Well, Google is remarkably coy about that. even the Google AdSense blog, which, inter alia, celebrates the success of partner sites, never actually states what the revenue each achieves is. Is it fear of being impolite, or of embarrassing the successful by showing just how successful they are, or is because Google is the only show in town that really succeeds in monetizing these sites - many of them falling into Chris Anderson's description of the long tail. 10,000,000 sites worldwide, generating just $50 per annum revenue for themselves would generate $500,000,000 revenue for Google if the search engine takes the same share of revenue as the sites themselves. If the revenue share is rather more skewed (another piece of data that isn't revealed by Google) and, say, 20% goes to the site and 80% goes to Google then Google's revenue from the same 10,000,000 sites, each receiving $50 pa, would be $2Bn.

It seems clear that the one real beneficiary of the monetization of the long tail is the one business that is truly pervasive.

20 January 2008