Goldfarb and Tucker (2011) use a large field experiment to separate the effect of matching an ad to website content from the effect of making the ad obtrusive.
Matching an ad to website content and making the ad more obtrusive each raise purchase intent on their own. Ads that do both do worse than ads that do only one. The penalty is strongest for people who refuse to give their income and in categories where privacy matters most.
Method
A large-scale field experiment on online display advertising. The design crosses contextual match with obtrusiveness and measures purchase intent. The authors also compare people who refuse to report income with people who report it, and compare categories where privacy matters more.
Main findings
- Each lever works alone. Content match and obtrusiveness independently increase purchase intent.
- Together they fail. Ads that are both matched to the site and obtrusive do worse at increasing purchase intent than ads that use only one of the two.
- Privacy is the suspected reason. The negative combination is strongest among people who refuse to report income and in categories where privacy matters most.
Why this is in the Darwin index
This is the field result on why a loud, tightly matched display ad can underperform a quieter match or a louder unmatched ad. It sits with the banner-avoidance papers as evidence that being noticed and being accepted are not the same outcome.
Source of record: doi:10.1287/mksc.1100.0583.