Commentary

What About Mobile Ad Targeting Inaccuracies?

by Laurie Sullivan , Staff Writer, February 29, 2012

While on a ride along in a helicopter, I witnessed a police officer reach for a smartphone and search for an image of a car -- make and model -- just seconds after dispatch provided a description of the vehicle fleeing a location where a shooting had occurred. The officer's partner needed a better visual description. The smartphone and the search engine provided it.

The shooting quickly turned into a murder, while the browser on the police officer's smartphone no doubt now has a cookie identifying the owner as someone searching for a specific make and model of a car. This is probably the only time the officer would want an image of that car. So I'm wondering: As more people use smartphones for business-related searches that might become inapplicable in the future, what percentage of audience segments is built from inaccurate search signals? Aside from skewed audience segments, will the percentage of irrelevant targeted ads rise as more people rely on smartphones?

If the officer had gone into Edmunds.com and searched on the term Chevy Malibu, for example, it might have become more indicative of intent, compared with someone searching on the keywords in a search engine, according to Simpli.fi CEO Frost Prioleau.

It's obvious that not everyone who searches on the term Chevy Malibu wants to buy the car. There are other products that might present a higher percentage of accuracy for ad targeting, such as searching for a flat-screen TV. Prioleau pointed to upper- and lower-funnel search terms. Certain keywords would have greater intent. Consumers will enter a site and price a car, but they are really not in market to make a purchase. Depending on the Web site, the search could drop the consumer into a false audience segment.

Simpli.fi, a DSP that supports both search engine and display marketing, launched a feature called Keyword Contextual Targeting on Tuesday. The targeting option enables marketers to use keywords to define, target, and optimize custom contextual categories for any campaign running on the company's platform. Brands can create custom contextual categories by uploading a list of keywords into Simpli.fi's platform.

Prioleau said this type of search retargeting for display ads helps to eliminate inaccuracies in targeting, but I want to know if others see problems that need to be worked out based on the rise in smartphone use. Perhaps this scenario contributes to the small percentage of inaccuracies the industry executives speak about.

display, search, search retargeting

2 comments about "What About Mobile Ad Targeting Inaccuracies?".

Check to receive email when comments are posted.

Eric Brown from Dataium, February 29, 2012 at 6:32 p.m.
Good story, this is why our firm uses Contextual Behavior in our analysis and data feeds. Our tracking would see this is the only car this user has searched within X period of time, or is an inventory search anomaly from his other car search activity, no keyword usage, websites visited, etc....all do not align with normal active in market auto shopper behavopr.
We have the consumers contextual behavior or behavior across thousands of automotive sites, or behavior beyond simply the last click to be much more relevant in predictive analytics, ad targeting and messaging.
Reply

Warren Lee from WHL Consulting, March 1, 2012 at 3:25 p.m.

Laurie, great topic and treatment. The inaccuracies in targeting can be addressed by leveraging other data sets. For instance Segmint analyzes spending data and patterns. Anomalies like the one you mentioned can be picked up and eliminated. Oh, and don't think that this is just a problem with mobile, as I think that similar inaccurate targeting takes place with all screens, certainly with TV.