'Pretargeting' Can't Work on Inferences Alone

Probabilistic Data Doesn't Go Far Enough in Targeting Consumers

By Published on .

Reprints Reprints

Despite advances in advertising technology, retargeting is far from an exact science. Ads that follow consumers who seem like good prospects are often actually hitting too late in the buying cycle, after the consumer has already made a purchase. (That's how ad tech often determines that you're a good prospect.) So retargeting can result in ineffective campaigns, often missing the mark on truly "in-market" shoppers.

To address these inefficiencies, the ad tech community has recently adopted the psychological concept of "priming," in order to "pretarget" audiences. Marketers can now screen audiences to find the consumers that are in-market and have declaratively expressed their purchase intent. The net result is custom audience segments of consumers that are ready -- or primed -- to buy.

Priming, a term from the field of psychology, refers to an implicit memory effect in which exposure to a stimulus influences an individual's behavior toward a later stimulus or event. The individual may be unaware that the first stimulus has influenced her response to the second stimulus.

Marketers who understand priming can use it to their advantage by pretargeting their audiences -- delivering appropriate ads to consumers who have already declared themselves in-market. By pretargeting, advertisers can capture consumers at the beginning of the purchase funnel. Pretargeting marks a vast improvement over plain vanilla retargeting, which often hits consumers after they've already made a purchase.

Tom Goodwin eloquently described the inefficiencies of retargeting in an Ad Age column. Goodwin notes that notwithstanding all of the advances made in personalization and data science, retargeting is still based on a murky collage of uncorrelated past behavior. He describes pretargeting as predictive advertising -- using data on various fronts to infer with good odds what consumers need, perhaps even before they realize it.

Unfortunately, simply using probabilistic data to pretarget will only enable marketers to hazard a guess that an ad will land in front of a consumer, either before or during the buying phase. While predictive advertising and probabilistic data are arguably more effective than most retargeting, they are still synthetic approximations for real answers.

By contrast, a deterministic data strategy has marketers asking consumers directly if they are, in fact, in-market for a certain item (for example, "Are you looking to buy a car in the next six months?"). Often this is accomplished through surveys, in a manner similar to attitudinal campaign research performed by ComScore, Nielsen Vizu or Dynamic Logic, where marketers create custom audience segments of in-market consumers. This flavor of deterministic data can sometimes be found in data management platforms.

Deterministic data can empower marketers to know with certainty that their ads are in front of the right consumers at the beginning of the purchase funnel. This level of confidence comes at a premium: It is more expensive than probabilistic solutions, but substantially more effective.

Marketers want to put the right ads in front of the right consumers at the right time. Advances in pretargeting make this achievable. The strategically powerful combination of pretargeting and deterministic data will change the way advertising is targeted in 2015 and beyond, resulting in better ROI on the buy side.

In this article:
Most Popular