Saturday, March 17, 2012

People Say They Object to Personalized Ads. Do They, Really?

A recent Pew Internet & American Life survey asked respondents their views of search engines and other websites collecting information about them and using it to either shape their search results or target advertising to them.

Overall, attitudes toward these practices are mixed, but the majority of Internet and search users express disapproval.

The study of 2,253 adults found that 73 per cent of search-engine users didn't want their search results to be tailored to them based on past searches, even when used only to personalize their future search results.

Over two thirds of respondents didn't want sites to do targeted advertising that uses their Internet history, and just over half said they had noticed targeted advertising that was clearly using such data.

Some 68 percent of respondents have an unfavorable view of the practice, saying they are not okay with targeted advertising because they do not like having their online behavior tracked and analyzed.

Some 28 percent said they are okay with targeted advertising because it means they see advertisements and get information about things they are really interested in.

That should come as no surprise. In the abstract, people also say they “hate” advertising. Only in context can actual user behavior be assessed. For example, though people dislike advertising, they are more than willing to accept it in return for something of value.

One suspects the same is true of attitudes about some specific cases of privacy versus personalization. In other words, a user with high privacy choices will not get personalized offers or relevant advertising. That isn’t to say users should be unconcerned about privacy breaches.

But the issue is that relevance, targeting and offers that are highly relevant to each particular person cannot be delivered without setting a lower bar for “privacy” settings. It’s always a trade off. So the issue is whether most users are willing to make trade offs of some privacy to get relevant, personalized offers and messages.

One suspects actual behavior would be different, when users understand they get something of value for sharing some information that might be objectionable in the abstract.

“I actually don’t think people are that concerned about security,” says John Basso, Amadeus Consulting CTO. “They want you to protect them.”

No comments:

Costs of Creating Machine Learning Models is Up Sharply

With the caveat that we must be careful about making linear extrapolations into the future, training costs of state-of-the-art AI models hav...