January 21, 2011

The Knights Templar Meet the Google Search Results Conspiracy Theories!

Blog Update (February 15, 2011): "The SEO Lament" - With Apologies to Gilbert and Sullivan

"When correctly viewed, everything is lewd. I could tell you things about Peter Pan, and the Wizard of Oz, there's a dirty old man!" -- Tom Lehrer, "Smut"

Greetings. I love conspiracy theories. I really do. I hardly ever believe them, but they're often great fun nonetheless, just through the sheer joyful lack of logic that usually pervades them to their inner cores.

My own theory has long been that most genuine conspiracies are the ones we don't even suspect exist. Once a would-be conspiracy has been labeled a conspiracy theory in public, it's usually already pretty much toast.

Yet people by and large love believing in conspiracy theories, for a variety of reasons -- for example, they can be very convenient for social control and personal gain.

In 1307, King Philip IV of France -- who was deeply in debt to the Knights Templar -- used the secrecy of their practices and rituals against them by declaring a vast Templar conspiracy -- triggering their mass arrest, torture, executions, and (conveniently) the erasure of Philips's associated debt. Even today, bizarre, ridiculous conspiracy theories regarding Freemasonry continue to circulate, egged on by the secrecy of Masonic rituals.

Conspiracy theories tend to fill information vacuums by attempting to postulate the inner workings of activities based on fragmentary and often statistically misleading observational data. This tendency is also exacerbated by the natural human desire to impose order on chaos, to assume connections where none actually exist, to try give ordered meaning to what otherwise might seem unpalatable asymmetric forces.

Even decades later, many persons refuse to accept the concept that one lone gunman assassinated President Kennedy, despite overwhelming evidence to that effect. It just seems so wrong that a single nobody with a rifle could change history so dramatically. Surely the existence of a vast conspiracy would make such events more emotionally tolerable, at least.

And so we come at last to the technology conspiracy theories of the moment, the increasing drumbeat of claims that Google is biasing their organic (natural, not paid-ad) search results in their favor.

The two sides of this current brouhaha may be best exemplified for now by long-time Google critic Benjamin Edelman's newly released Measuring Bias in "Organic" Web Search, and search marketing specialist Danny Sullivan's response in Study: Google “Favors” Itself Only 19% Of The Time.

I must admit to being somewhat amused both by Benjamin's problematic statistics and by Danny's detailed analysis of those numbers. In particular, the former's methodology is so highly questionable on its face that Danny's effort in this case strikes me as being somewhat akin to publishing a deep, intellectual analysis of a brief Monty Python skit -- the source materials are much more appropriately viewed strictly for their comedic value (I know a dead parrot when I see one!)

In a more serious vein, we note that statistics, though seemingly composed of hard facts and numbers, are in reality almost infinitely mutable, as are the studies that quote them. Darrell Huff's wonderful 1954 expose and guide, How to Lie With Statistics (very much still in print!) is as marvelously relevant now as it was more than a half century ago.

Of course there's nothing funny about the ramifications of bias accusations against Google. And as in the case with King Philip and the Templars, it's easy to find financial motivations in play among some of Google's accusers -- who often tend to be allied with Google's competitors either directly or via "astroturf" relationships.

To be sure, the detailed mechanisms of search ranking algorithms (which in Google's case undergo virtually continuous tweaking for anti-spam and relevancy adjustments) are quite opaque to outside observers.

But it's hard to see how this could reasonably be otherwise, without opening up search results rankings to endless "gaming" of results by spammers and other bad actors, to the vast detriment of Internet users generally.

And after all, search results, whether by Google, Bing, or anyone else, are merely the opinions of those search services at any given moment in time -- not some sort of commandments handed down on stone tablets.

But ultimately what decimates the Google Search Results conspiracy theories for me is that they just don't make logical sense.

Just as Tom Lehrer noted in his classic ditty, given a suitable observational bias, even Peter Pan can be viewed as obscene. But is such a bias reasonable? Of course not.

Similarly, in order to assume purposeful organic search results bias by Google, we need to also assume either (a) that there's a logical upside to doing this for Google, that significantly exceeds the downside risks of being caught doing so, or (b) that Google management would behave in an unethical, potentially self-destructive manner for no logical purpose.

To accept either of these assumptions seems nonsensical. I don't believe Google operates unethically or illogically.

But beyond this, Google simply doesn't need to bias natural search results. They are incredibly successful by any measure, and any possible incremental advantage to manipulating organic results would be enormously swamped by the public relations and other risks that could result from their doing so.

What's more, while it's likely true that most people tend to click initially on highly ranked results, it's also clearly the case that most Internet users are not wearing blinders that restrict them only to the first few results.

If users are not satisfied with what they find initially, they'll usually be back looking at more links and trying out other services in short order. Most Internet users are simply not the search result automatons that some studies would seem to presuppose.

The attacks on Google's natural search results rankings have all the hallmarks of classic, opportunistic conspiracy theories. Not only do they not make sense "by the numbers" -- but they also are illogical when viewed more broadly in terms of Google cost/risk/benefit analysis.

As far as I'm concerned, Peter Pan's small friend Tinker Bell is more likely to exist, than for the search results manipulation bias accusations against Google to be accurate.

And that's the case even if you don't believe in fairies.


Blog Update (February 15, 2011): "The SEO Lament" - With Apologies to Gilbert and Sullivan

Posted by Lauren at January 21, 2011 12:06 PM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein