In yesterday’s post on “The Great Inventory Split,” I talked about the very real and growing bifurcation between vast amounts of low-quality, low value content and focused pools of high-quality, high-value content available to online advertisers. For the last day or two I’ve been attending Google I/0, the company’s developer conference, in San Francisco. As I type, I am listening to Google’s Marissa Mayer, VP of Search Projects and User Experience, describe the several of the problems that Google is working on. General search, of course, tops the list. Unlike other companies, like Barry Diller’s IAC/Interactive, which is working to create segmented, vertical search experiences like Rushmore Drive, Google continues to try and tackle the “impossible problem” of parsing single, general queries to determine semantic intent, local application, automated cross-language results, etc. Marissa is continuing to run down a laundry list of cool, math-camp kind of projects that Googlers work on—often without real concern for whether a project can become a business, but instead for whether they could solve a problem, explore a concept, or push an idea to its limit. StreetView emerged from one of these projects. Google Book Search is another. Her presentation is clearly a marketing effort (for their “20% time” idea, which I don’t need to explain) versus 99% of the rest of the conference which is fairly hardcore, and very good.
Thinking about yesterday’s post and Marissa's talk, I’m seeing an interesting relationship between the two general categories of publishing inventory on one hand, and the two broad approaches to search we’re seeing emerge (verticalized/segmented vs. general). In both cases, the category where the goal is to succeed against a broad, general challenge is the harder one. In both ad monetization and search, the challenge is about zeroing in on meaning without taking the starting-point shortcut of beginning with a smaller content pool. Think of it as the needle in the haystack problem: if you are able to start with a smaller haystack, you’re better off, but the real feat would be to find the needle in the hayfield. Google wants to solve the “needle in the hayfield” problem. Why? Because at the moment, working on the hard problem doesn’t hurt Google’s business, and it keeps its engineers engaged and motivated. For others, who lack the huge market share and corresponding ad revenue, and whose sole focus is to shift that market share and revenue, different guiding principles may apply.
Really, though, this shouldn’t be a surprise, as it is universal: formulating a general theory of relativity continues to prove a challenge even after Einstein’s (generally successful) crack at it, while event prediction in closely controlled environments is commonplace.
As a final note, I happened start reading The Black Swan by Nassim Nicholas Taleb on my flight to Google I/O. One of the advantages of being part of a positive Black Swan event is that one is afforded the luxury of being able to choose one’s next pursuits purely on the basis of one’s own interests, rather than the market's (or one's investors'). As Taleb notes in one anecdote, for a trader capitalizing positively on a Black Swan event, one good year out of ten, or even one out of one hundred, is enough. Clearly, Google’s actions continue to follow this pattern generally: so far, they have had the equivalent of “one good year” – but that one may be enough to cover them for a long, long time.
