Behavioral advertising ≠ government surveillance. It's time to clarify
Recently, issues of data privacy and state invasion of privacy have not left the pages of newspapers and blogs. Since this was a key issue for us from a business point of view, we had to tell and prove regularly in professional meetings and in private conversations, how careful use of data for specific business purposes differs from tracking Internet users. Therefore, the article, the translation of which is given below, seemed to me very relevant.
When we started developing algorithms for optimizing online advertising in Maxifier, the issue of data accessibility and confidentiality was one of the first that depended on the potential capabilities of our product and the general positioning of the company. For ourselves, we unequivocally decided that we would not use personalized user data, but only work with averaged statistical data. Yes, this is somewhat poorer in algorithms. Yes, the result for each individual person will be slightly less accurate - because we will use automatically reconstructed patterns of behavior, but not knowledge of the behavior of this particular user on the network. Created a self-learning system; over time, it becomes more and more “smart” - statistical knowledge about users is accumulating, templates are being refined. This solution allowed us to avoid all the problems with the subsequent data privacy laws in Europe, for example, which then affected all behavior targeting systems. And the fact that we initially built the algorithms, assuming a high level of noise and data coarsening, later we even allowed us to make this an additional competitive advantage - because our product works in cases of lack of information or lack of access to it due to the policies of sites and ad networks - in situations where the decisions of competitors-owners of databases could not cope with quality indicators. assuming a high level of noise and data coarsening, in the future we even allowed us to make this an additional competitive advantage - because our product works in cases of lack of information or lack of access to it due to the policies of sites and ad networks - in situations where the decisions of competitors-owners of databases could not cope with quality indicators. assuming a high level of noise and data coarsening, in the future we even allowed us to make this an additional competitive advantage - because our product works in cases of lack of information or lack of access to it due to the policies of sites and ad networks - in situations where the decisions of competitors-owners of databases could not cope with quality indicators.
“Snowden's Summer” has died down, and articles that draw parallels between collecting data for online advertising and government surveillance are increasingly catching the eye. National Public Radio, the largest state radio station in the United States, went even further: it aired a story that presented these two practices almost interchangeably, and the New York Times casually referred to a collection of personal information collected by the US National Security Agency (NSA), in the text of an article about advertisers using mobile call data.
Confusing two so different types of data in the minds of readers is, at least, irresponsible. The government collects information regardless of the desire of its owners. It can creep into any corner of our lives: from mobile calls to chat messages and search queries, and it will find many pretexts to circumvent the rules. But online advertising, in order to get your data, requires permission from at least one of the parties involved in the process, either the consumer or the publisher’s site. In addition, the European Union should notify consumers that their data has become available to advertisers . Stories that identify the actions of government and digital marketers will hamper innovation than not only make advertising less relevant and content creation less profitablebut will also scare people away from participating in online healthcare , education, and transportation (pdf) services if they request personal information. These features are based on users trusting new products requiring the exchange of personal information.
To understand how big the difference is between collecting data for online advertising and government intelligence, look at the methods for obtaining and using it. If the United States government wants to get information about you, it will either use technology such as PRISM or use the Foreign Intelligence Surveillance Act to force companies to transmit data and prevent them from informing you of the leak. In some cases, information goes toThe Drug Enforcement Administration (DEA) and the IRS , from where it is transferred to local law enforcement agencies, where they can use it as much as they need to come up with a story that will cause your arrest and hush up the true source of the data. The New York Times reports that this data is used to combat anything from cigarette promotion to copyright infringement - although such incidents rarely threaten national security.
As a rule, in the advertising business, data is collected in order to sell the maximum amount of goods, spending a minimum of marketing resources. The data used includes a list of search queries, the web pages you open, things that you viewed in online stores, and recently, information about exactly what you bought. In the West, there is strict regulation that requires the rejection of personalized information and prohibits the use of any data to make a decision, for example, about the amount of insurance payments to the user. The success of advertising based on the management of large amounts of data is determined by the reduction in the cost of its placement, and the consumer ultimately sees a more relevant advertisement.
Saying that the government can use information from advertisers for espionage is a clear exaggeration.
The NSA collects data on user actions from several sources , and they are much more personalized than those used to target ads - so the existence of a behavioral aspect in the advertising industry does not make it easier to collect information for the NSA.
The New York Times Editorial Board attempted to construct a situation in which there was a need to regulate behavioral advertising, and came up with a hypothetical scenario for how data collected for advertising could harm: for example, information is stored in personal profiles, and government structures may require access to them at any time. Or they can be bought by employers or banks to evaluate candidates for employment and granting a loan.
However, this hypothesis is unfounded. The United States Law on The Fair Credit Reporting Act requires a high degree of transparency of any data used to make a decision on financial soundness. The government does not need the data collected by advertising, because the data of the NSA are obviously more accurate than those used for targeting. In a recent Time magazine reportit is said that the NSA will get a little less than 500,000 address books per day. Obviously, the existence of the behavioral advertising industry is not able to further simplify the collection of information for the NSA.
In fairness, we note that there are examples of the dubious use of marketing data. Last year, the Wall Street Journal recorded several cases where cookies were used as a tool to change pricing : customers were offered different prices depending on their territorial proximity to competitor stores. We also talked about cases of increase in airline tariffs after repeated searches. But even this is not the most painful; the feeling of "someone watching me" is much more depressing.
But the massive attacks on behavioral advertising undermine online models without confirming the key assumption: are consumers really worried about using data for advertising if they correctly understand how they manage their information? To find out, we turned to Enliken research: according to its results, less than 10% of users are worried about the use of their data for marketing purposes. Acxiom analysts came to the same result when they opened their databases to the public a couple of weeks ago. Research by PricewaterhouseCoopers (pdf) , DMA, McCann (pdf) confirms the idea that consumers want to share information with brands they trust.
Therefore, all types of online businesses are interested in transparency and regulation of the processes of working with private data. Knowledge of the preferences of a particular consumer allows you to raise services to a fundamentally different level. It would be a shame to lose all the opportunities that open before people because of the banal confusion of concepts. Let's separate the flies from the cutlets.