The Cost of Search: Bing Edition only $2.6 Billion
July 26, 2011
In July 2008, I wrote “Yahoo Cost Estimate”. Here’s the key passage:
I wanted to run through some of the cost data I have gathered over the years. The reason is this sentence in Miguel Helft’s “Yahoo Is Inviting Partners to Build on Its Search Power,” an essay that appeared in the Kentucky edition of the New York Times, July 10, 2008, page C5: “Yahoo estimates that it would cost $300 million to build a search service from scratch.”
At the time, the estimate was what I call a “crazy number.” How crazy was Yahoo’s guess-timate? According to “At Microsoft, Bing Too Costly to Keep,” on July 25, 2011, New York Times, section B 2 and in the July 24, 2011 article “Bing Becomes a Distraction for Microsoft”, we get some numbers. Yahoo couldn’t do search and it owned Inktomi and Fast Search’s Web search system. In 2008, Yahoo had $300 million and couldn’t do search then. Today Yahoo does not do search. Yahoo recycles Bing results. Now Bing is, if the New York Times is correct, too costly for Microsoft. Yikes. Search is expensive.
How does this sound? “Bing lost $2.6 billion in the latest fiscal year.” The Yahoo estimate was off by a factor of eight. Give Yahoo a break for inflation over the last 36 months, and the Yahoo number was wrong by a little less. On the other hand, when one adds up the total costs of Bing, the cost of Web search reaches Greek debt-scale levels.
What’s with the cost of Web search?
I have written extensively about the cost of search. You can get a run down of the various monographs which contain my thoughts about enterprise and Web search costs by running a query on this blog or looking at one of my books, monographs, and journal articles which explore search costs.
Here’s a rundown of the challenges.
First, on the Web information expands rapidly. Within the last three weeks, Google+ ran out of space. Search is not really tightly integrated with that service. If Google has problems scaling, other companies will too. And scaling is a black hole of search costs. As information expands, the infrastructure must be able to accommodate the data and the outputs of the indexing process. So lots of information to process equal lots of costs. In Microsoft’s case, we have a number which may or may not be accurate. But $2.6 billion is interesting. The New York Times does some fancy dancing around the $2.6 billion, but with the Web search landscape reduced pretty much to a handful of companies, the costs are a factor.
Second, I think that digital content is growing rapidly. No one knows exactly how fast, but I know from my test queries that certain content is either not indexed or it is no longer available. I think that indexing systems are becoming more selective and organizations, particularly government agencies, don’t have the money to make the information available as in the past. The result is that we may never know how much digital information is “on the Web” and it’s a safe bet that finding non indexed content is going to be more and more difficult. As a result, costs go up for what’s there, and there is neither money nor appetite for what is not indexed. Will commercial database producers pick up the slack? Nope. Too expensive. So the free Web index ride is over. Commercial services won’t move in. The economics of everything related to low value information mean content is going to be less findable. Maybe a breakthrough will reverse this situation, but I predict a worsening of content access for the foreseeable future.
Third, users want systems which think for them. The costs of transforming content and then using next generation methods to make that content via “search without search” methods are costly to “invent” and costly to maintain. As a result, the information which becomes available will be content that can be easily monetized. The only way to make the numbers work is to focus on what sells and then find a way to monetize that content. Getting the “what sells” part wrong will sink a search engine before the ink is dry on the VC’s first check. Getting the “monetize” part wrong means the company will be shut down, probably more quickly than in the past.
To sum up, the costs of search are interwoven. There’s the plumbing cost, the technology cost, the marketing cost, the sales cost, and the opportunity cost. When Microsoft cannot afford search, who can? Right now, there are just a few answers. Do you use Google, Blekko, Exalead Web Search, Yandex, Baidu, Jike, or Bing? How many will be financially viable in one year? History teaches us that there will be attrition in this cost battle.
And if information is indeed infinite, won’t the cost be infinite too or at least $2.6 billion?
Stephen E Arnold, July 26, 2011
Sponsored by Pandia.com, publishers of The New Landscape of Enterprise Search
Comments
-
- Subscribe to Beyond Search
Feature archive
News archive
-
