Thinking Out Loud: Relevant or Wrong

[Is targeted advertising a boon or a privacy violation waiting to happen?]

Behavioral targeting (BT) is the new snake oil, a veritable cure-all for the online marketing industry. There is a difference, however, between BT and 19th-century snake oil; those touting BT actually believe they have a magic elixir, while the itinerant salesman in the Conestoga wagon knew full well that he was fleecing the unsuspecting and gullible town folk.

Upon closer examination, the rancid formula of this “BT elixir” should give any and all in the media food chain pause for concern before ingesting. And in the long list of what's wrong with targeting ads by behavior, three elements stand out as particularly toxic.

Let's use the FTC's definition of behavioral targeting: “… the tracking of a consumer's activities online - including the searches the consumer has conducted, Web pages visited and content viewed - in order to deliver advertising targeted to the individual consumer's interests.”

BT is an invasion of our privacy, for even if we opt in to the process, we're often exposing personal data then obtainable by the government via subpoena. The purposely stealth nature of this pseudo- science evokes chilling, Big Brother-esque imagery, and defies the notion that something of such questionable moral foundation could proceed this far unchallenged.

The fact that private industry now works in tandem with the government to stalk and track us online may deflect but cannot dilute the moral argument. It seems little considerations like the prohibitions against “unreasonable search and seizure” enshrined in our Bill of Rights never made it onto the BT agenda.

BAD PREDICTORS

Besides being creepy, targeting relies on spurious predictors. Once, Wall Street hired thousands of learned mathematicians to portend the future of the financial markets. They poured over millions of transactions and developed sophisticated algorithms to predict market trends. Science was heralded as a welcome savior capable of taming the chaos of the stock market. The underlying principles of the mathematics were developed by John Nash, who found that there was an underlying predictive order to minute human interactions.

Amass enough data, they said, and the Ph.D.s could ferret out the predictive variables that made people do what they do.

Only one problem: They were all wrong.

None of the models that identified predictive variables envisioned the bus careening off the cliff because all of these models were conceived through the rearview mirror. The flawed variables derived from past, irrational behavior did not (and could not) predict the devastating financial collapse that ensued.

Thousands of suddenly unemployed Ph.D.s left Wall Street. Where did they end up? Next stop: Madison Avenue, Google and BT shops.

The underlying mathematical ingredient of BT is snake oil, but with “a spoonful of sugar to help the medicine go down.” (For a thorough examination of the mathematical nonsense, see Nicholas Taleb's two books, Fooled by Randomness and The Black Swan.)

As instability increases, mathematical reductionism renders predictions of future behavior untenable. Simply put, driving by looking in the rearview mirror is a sure recipe for disaster.

RELEVANCE IRRELEVANT

The main ingredient in this serpentine concoction is relevancy, the composition of which comprises “getting the right ad in front of the right person at the right time.” It sounds so logical and reasonable.

Of course, witch trials were also considered logical and reasonable once upon a time.

What the relevancy ingredient assumes is that people want relevant ads. In truth, the data suggests just the opposite. They don't want ANY ads at all! But by asking questions in a subjective, self-serving manner, you can pervert the Socratic Method to get the answer you want.

Former FBI director Louis Freeh said: “Ask the American public if they want an FBI wiretap and they'll say, ‘no.' If you ask them do they want a feature on their phone that helps the FBI find their missing child they'll say, ‘Yes.'”

If audiences are asked, “Do you want more ads?” all data suggest they will answer with a resounding NO! But if they are first asked, “Do you prefer relevant ads over irrelevant ads?” they may reasonably, albeit reluctantly, say yes. Ask a stupid question, get a stupid answer.

In an on-demand world, nobody demands more advertising. An audience so empowered makes advertising a challenging endeavor, especially when all this snake oil leaves such a bad taste in our mouths.