Here is an interesting paper on research into utilizing predictive analytics to predict the likelihood of someone bouncing from an online ad (i.e. specifically Google Adwords). The practical definition of bounce is someone clicking on your ad, coming to your site, and leaving pretty much immediately. That is to say, they took one look, were not impressed (generally because they found it irrelevant to what they were looking for or did not like the look of your site).
The researchers found that predictive analytics could be used to increase the ability to predict the likelihood of someone bouncing. To quote an excerpt from the conclusion "We have also shown that even in the absense of substantial clickthrough data, bounce rate may be estimated through machine learning when applied to features extracted from sponsored search advertisements and their landing pages"
The full paper written by D. Sculley, Robert Malkin, Sugato Basu and Roberto Bayardo can be found here.
No comments:
Post a Comment