The Federal Trade Commission (“FTC”) hosted a panel discussion, in late March on “Alternative Scoring Products” as part its 2014 Spring Privacy Series, signaling the Commission’s increased attention on this burgeoning industry. The FTC has indicated that its “goal is to study what is happening in the alternative scoring space, what may be on the horizon and what potential privacy concerns these products may raise”. Alternative scoring products use consumer data to predict future products or services in which consumers may be interested. Predictive modeling is based on the concept that algorithms learn from data. While the FTC itself did not offer any guidance as to particular legislative or enforcement courses of action, the panel’s focus on privacy concerns further reinforces the idea that big data is an industry in which the FTC has shown increasing interest.  Representatives from public interest groups voiced concerns over the transparency of the data models used in alternative scoring products, while those from the data analytics industry reasoned that current privacy regulations and industry self-regulation provide the necessary oversight for data analytics.  All of the panelists agreed that the use of alternative scoring products was here to stay, and will only become more prevalent in the future.

Public Interest Groups Want to “Peek in the Black Box”

One of the main concerns voiced by public interest groups was the lack of transparency of data used in data analytics models, and in alternative scoring technology in particular.  Greater transparency, they reasoned, would demonstrate to the public and regulators that companies do not use alternative scoring in a discriminatory manner.  While consumers may prefer advertisements that are more relevant to their needs and desires, public interest groups worry that predictive data analytic models may limit offers and opportunities to certain consumers due to the data used in those predictive models. Representatives from public interest groups contrasted new alternative scoring methods used by companies to the long established credit score.  The representatives cautioned that while the Fair Credit Reporting Act (“FCRA”) prohibits the use of discriminatory factors – i.e. race, religion, gender – in computing consumer credit scores, alternative scoring products may not fall under the law’s purview, and escape regulation to the detriment of consumers. Consumer advocates were also concerned with the use of aggregate scores – information gathered on individuals from the communities in which they reside– to potentially deny opportunities to individuals based on the characteristics of their surroundings instead of individual characteristics.

The panel discussed the effects of predictive scoring by debating the findings of a Wall Street Journal investigation that studied whether online retailers varied offers or product pricing based on consumer information extracted from online profiles.  Some of the elements contained in the online profiles included web browser type, user location, and browsing history, amongst others.  The study found that certain retailers offered different prices for the same product based on the location of the consumer, and that some credit card companies presented different card offers based on differing user profiles.  Representatives from the data analytics industry countered that several factors affect pricing, including customer loyalty, which can decrease the prices paid by frequent customers.

Industry Groups Tout the Pro-Competitive Benefits of Predictive Analytics

Representatives of the data analytics community emphasized the ways in which predictive analytic tools can be used as a force for inclusion of consumers in the marketplace, rather than exclusion.  Panelists promoted the industry’s use of multiple sources of consumer data to better target consumers based on specific characteristics, leading to better consumer offers and increased competition to the benefit of consumers.  They also discussed how new data analytic technology can help detect fraud at an early stage by verifying the device a consumer uses in a transaction as his or her own.

Downplaying the discriminatory risks associated with alternative scoring products, industry representatives argued that the technology is used only for advertising purposes, and is not used to determine whether a consumer qualifies for a particular product. They also noted that many existing laws, including FCRA, the Health Insurance Portability and Accountability Act (“HIPAA”) and the Graham-Leach-Bliley Act (“GLBA”), already regulate the field of predictive analytics, and argued for greater industry self-regulation, as opposed to the creation of new legislation.

A Look towards the Future of Alternative Scoring Products and Best Practices

Despite the panelists’ diverging views, one thing on which all agreed is that alternative scoring products are here to stay.  The tension between the convenience of targeted ads and the risk to consumer privacy remains, and is certainly on the FTC’s radar.  In a time of increased scrutiny on these tools, panelists proposed solutions and best practices:

  • Ensure existing predictive analytic tools are in compliance with existing privacy law frameworks.
  • Be especially cautious about the use of consumer data relating to health, finances, and children.
  • Create a privacy policy that is explicit and transparent in the ways in which consumer data will be used.
  • Be aware of the types of consumer data your company collects and ensure that data is not used to discriminate against certain groups of consumers.