I recently came across a couple of smart observations on the diminishing quality of Google search results (examples here and here) that nicely crystallized my emerging, vague sense of displeasure and frustration with my default gateway to the internet. Essentially, the trouble is that there now exists decent (and in some cases, excellent) and widely-available technology for persuading Google’s algorithms of the relevance of one’s content; the banal jargon for this is ‘search-engine optimization’ (SEO).
I view SEO itself as a neutral instrument, with good and bad potential applications. As a first pass, I’ll define ‘good’ and ‘bad’ roughly in terms of the decrease or increase in my transaction costs of finding information. Suppose I’ve just written a great book on the history of some obscure typeface. If I have strong belief in the merit of my content, perhaps I’ll be willing to invest in some extra help to beef up the visibility of my web presence among typeface enthusiasts, in the hope of selling more books. I may even need to share some of this content for free, to help potential customers (not to mention the algorithms!) evaluate my credibility. In general, everybody wins: I sell more books, the SEO company makes money, Google delivers the relevant search results that keep you coming back, and the masses find useful information more easily. This sort of theory should also bias Google’s AdWords model in favor of relevance. If I’m willing to pay a hefty price per click for some heavily-searched term, I ought to be pretty confident I can monetize those clicks by delivering something relevant to searchers. Google became dominant largely because it was the first and best to win the war for relevance, and it managed to figure out a way to make money in a way that reinforces that relevance.
Of course, even a well-designed ecosystem is vulnerable to spam. That vulnerability increases in systems where there is a lot of potential money at stake and there are trivial transaction costs and no barriers to entry, since even a microscopic success rate can yield a positive return on investment. Advertising dollars have been migrating towards the internet for years and it is possible to earn money as a ‘distributor’ for Google and others, i.e., by hosting ads that compensate webmasters on a pay-per-click basis (as opposed to the model of a decade ago that might have required selling banner space). At the same time, the cost of creating a web presence has continuously fallen, as bandwidth, storage, domain name generation, and the like, have become commoditized and in vast supply (relative to the size of an individual page). And, crucially, the exponential growth of user-generated content on all manner of topics (from Wikipedia to blogs and message boards) has even reduced the last barrier to entry, i.e., actually having some content.
Spam on Google looks something like this: I scrape a bunch of content and links based on some keyword, surround it with some boilerplate, toss on a couple of stock images, and call myself an aggregator (who has really done nothing more than, well, run a Google search for you). I am positive you’ve seen something like this already. You stumble upon my page by some means, realize you’re stuck in something useless, and probably go back to Google. But there’s a tiny chance you might click on one of my links, and there’s a tiny chance that it might be an advertisement where I receive a per-click share of revenue. If my costs are low enough, the business model works. SEO enters the equation by helping me get to the top of your result page – essentially, sending the spam to your inbox – and, to a lesser degree, helping me increase the probability that you will actually click on an advertisement in my page.
I think Google is still pretty excellent at finding very specific pieces of content that I’m already aware of. For example, I recalled reading about a set of predictions from the New York Times in 1931, and Google helped me painlessly track down that exact piece. It seems doubtful that one could reliably spam my Google search for ‘new york times article about predictions from 1931’. But Google does not really monetize its edge in delivering this kind of content, either.
Google no doubt has armies of rocket scientists who spend their time thinking about this issue in terms I can’t comprehend. But my point here (about which I don’t claim originality of authorship) is that a variety of trends are converging to increase the viability of ‘Google spam’ as a business model in ways that weren’t true a year ago. This will continue to weigh on the debate about ‘algorithmic’ vs. ‘curated’ information, which I may take up in the future.
P.S. I hope most readers will not understand the title of this post and will assume it’s a shameless attempt to boost my own search rankings on the coattails of a Real Housewife. Because that means I can rock your world with this.