MIT: Personalized Ads Don't Always Work


Personalized advertising isn't all it's been cracked up to be, according to a new study from MIT.

Contrary to popular practice, personalized ads not only don't drive conversions, but are likely to be ignored, according to the study by MIT Sloan School of Management Prof. Catherine Tucker and London Business School Prof. Anja Lambrecht.

Advertisers now have the ability to collect huge amounts of data about each individual, and it has always been a "truism" that personalization increases relevance which therefore increases conversions, Tucker says.

"However, it may not be effective to use every piece of datum when designing ad-content," Tucker tells Marketing Daily. "For example, when an advertiser observes me abandoning a shopping basket with a pair of sneakers in it, it does not mean that it is necessarily effective to use that information to showcase sneakers in every subsequent banner ad I see for the firm. We find, instead, that such a "specific" form of ad content works well only when there is external evidence that people are close to making a purchase."



The data for the project were provided by Havas Digital from the Artemis Database. "They were excited with the results because they wanted insights about how to better target customers using different types of advertising along the purchase funnel," Tucker says.

Using data from an online travel firm, Tucker and Lambrecht addressed the questions of whether it is always optimal for advertisers to provide more specific ad content based on consumers' earlier product interests, as well as when increased specificity of advertising is effective. In their study, consumers who had looked at a travel company's specific products were randomly shown either a generic or a personalized ad. The company then tracked whether they ultimately purchased one of its travel products.

People who saw the personalized product information were on average less likely to buy a travel product on the day they were shown the ad than the people who were shown the more generic ad. Highly specific ads were generally not more effective than generic messages.

However, Tucker and Lambrecht found that this was not always the case. The key, they said, was whether the consumer had developed well-defined product preferences. When online shoppers were simply looking at a product category, ads that matched their prior Web browsing interests were ineffective. However, after consumers had visited a review site to seek out information about product details -- and were closer to a purchase -- then personalized ads became more effective than generic ads intended for a mass audience.

As a result, notes Lambrecht, marketers should carefully evaluate when to use new online advertising techniques, paying attention to how they can identify whether a consumer thinks broadly about a product category or seeks out detailed information.

In the study, consumers who had visited travel review sites to compare hotels reacted more positively to an ad that gave them more information about those hotels. But if they were just browsing, then they were more responsive to generic messages.

This was a "big surprise," says Tucker, as there has been a lot of excitement about using this new technique in online advertising where marketers can reach consumers with messages that are a better fit based on their known interests. "But it turns out that just because you have the data to personalize, it doesn't mean you always should," she says.

2 comments about "MIT: Personalized Ads Don't Always Work ".
Check to receive email when comments are posted.
  1. Michael Kelley from AdGenesis, June 3, 2011 at 3:12 p.m.

    While I appreciate the significant effort to study the impact of ad personalization on conversion and attention, AdGenesis is seeing a completely different set of positive results on both fronts. Perhaps, the definition of “personalization” is either to broad or misused in the study. For sake of a constructive response, let’s go with the former. The study looked at personalization based on a consumer’s past behavior to detect inclination to buy and when best to deliver a relevant ad. The study showed past behavior is not a good indicator of personalized ad performance. We concur. At AdGenesis, we look at personalization based on how a consumer currently lives and their upcoming purchase intentions. Working through our publishing partners, we deliver a private label solution that allows publishers to match consumers to brands and drive more value for all. So, how does AdGenesis match branded video messages to consumers? We simply ask. By asking consumers how they live and what’s on their(upcoming) shopping list be it vacations, shoes, cars, phones or computers, we can better deliver video messages of relevance to them. This value exchange -- along with offers from the brands--produces results that include more than a 30% open rate of emails/SMS messages alerting users of brand matches, 98% proven video view rate with our attention guaranteed technology and 11% or more click-thru rate. As such, we completely agree with the researchers’ conclusion: “when online shoppers were simply looking at a product category, ads that matched their prior Web browsing interest were ineffective.” This is why we created AdGenesis for brands, publishers and consumers. We are now glad to see the evidence to support our definition of “personalization” as well as our business model. Michael Kelley, CMO,

  2. St.jon Clark from PortableIQ, June 16, 2011 at 5:33 p.m.

    Interesting study and comment Michael Kelley. As an ad savvy marketeer I know the only ads I consistently click on are the ones that target my passions. It's like I can't help myself. Since those have plainly been identified, that type of personalized advertising is working. It's clearly based on monitoring the content of the sites I spend the most time on and visit the most frequency. What's old is new again. Content is King.

Next story loading loading..