Data mining in the media economy

It seems things always come in threes. This also holds true in Privacy News. Over the past holiday week, the US saw a series of pieces related to the on-going privacy/targeting debate in the media.  Stories percolated into Business Week and The New York Times related to the regulation/self-regulation discussion of behavioral targeting, while in the UK it was announced that British Telecom was terminating its controversial trials of the Phorm targeting service. What’s different this week? Just maybe the industry trade bodies are finally taking seriously the threat of external regulation? With that there is an escalated need to aggressively pursue a self-regulatory path.

A cross-section of industry groups have now produced 7 guiding principles for behavioral targeting and taken a passable stab at defining it:
“The collection of data online from a particular computer or device regarding Web viewing behaviors over time and across non-affiliate Web sites for the purpose of using such data to predict user preferences or interests to deliver advertising to that computer or device based on the preferences or interests inferred from such Web viewing behaviors. The Principles do not apply to a Web site’s collection of viewing behavior solely for its own uses” ¹

The specificity of the definition is very interesting in that it draws some very specific data practices into question. These relate to specialist companies that aggregate data across sites such as NextAction, Acerno, and data aggregators such as Blue Kai. However it may exclude companies such as Media6Degrees, that makes inferences based on your association to users you are connected to via social media. While such companies are very familiar to interactive advertisers, they are practically anonymous to the consumer and are, as such, clearly in the target area of regulators because they hold an ambiguous and undefined role.  Do we want to go this route and call out companies?

How does this apply and does it also challenge re-targeting and other practices related to data collection?  Re-targeting shares data from advertiser sites with media networks to allow the re-solicitation of visitors to an advertiser site.  You may have experienced this when you visit a site and then almost immediately start to receive ads from that company. A recent article by Wendy Davis in Mediapost shed light on this practice when she reviewed AT&T’s data sharing activities. Greg Neal , VP of Product Strategy at my company, TruEffect, observed in his analysis of the top 500 advertisers that, on average, data is shared through cookies 7-20 times, depending upon the industry.  

“AT&T is by no means unique" said Neal. "In fact they are better than many and have a very clearly articulated privacy statement related to these practices.  We typically see data being shared with between 3 and 10 companies.” What is more concerning for Neal is companies that do not understand how their sharing of data impacts their business and the brands they support.

“When reviewing client data practices it is not unusual to see data they have shared for specific re-targeting to be driving advertisements for competitive advertisers. I think advertisers often over-estimate the control they have of their data. When they share that data with another party, there is little control or regulation over the specific purposes and re-purposes for which the data may be used. It’s also extremely difficult to monitor or track. When the data is gone, it’s gone!”

This is a view shared by Match.com’s Jim McDonald who said, “We have rationalized the range of networks with whom we work based on compliance to our data standards. We have also begun investigating other methodologies such as managing all media within our domain. Users have a relationship with Match not any third party company they’ve never heard of.” Research supports this view with 61% of users expecting their interactions to be private, but nearly the same number also want advertising to be relevant.

Chris Ahrens of agency Draft/FCB reiterated this sentiment saying, “Draft/FCB has taken a pragmatic approach to the BT area and deployed different technologies when most appropriate.  We were the first to use ad serving without cookies as well as supporting the use of the advertiser option when appropriate”.

Given advertisers have options, is the industry’s response really getting ahead of the issue or simply acknowledging a changed reality?  Behavioral Targeting comprises less than 9.5% of all display media sales² although they do comprise a much larger part of media selection and retargeting.

Nielsen’s Marc Ryan believes that many consumers have already created their own opt-outs. “For quite some time we have been noticing an increasing trend in users managing their data more aggressively, this includes deleting or simply not accepting cookies. Typically we see more than 50% of media being managed in this way by better tools and options in the browser and use of security software.”

Webtrends, a leading web analytics company, saw similar trends.  “We believe about 28%-33% of users now manage how their data is used in media.  Consumers are ok with sharing data with companies they know (and trust).  They are often a little more savvy than the industry gives them credit for or wants to admit as it impacts measurement accuracy and performance.  In site-side analytics we saw this a number of years ago and made adjustments by working in the domain of the advertiser. We see the same trend in media and believe companies such as TruEffect offers solid alternatives to current practices.”

– Martin Smith
¹ Self Regulatory Principles for Online Behavioral Advertising
² eMarketer, June 2008

Comments

[Data recovery]

Data mining is the process of extracting hidden patterns from data. As more data is gathered, with the amount of data doubling every three years

Thanks & Regards
Harleen
System Analyst
http://www.recoverybull.com

[Dentist in Blackpool]

Data importance explained nicely with useful information and interesting concepts.

[Economic Crisis]

Very good and informative article about data mining in the media economy