PASS Business Analytics Conference Day 2 Keynote: A Framework for Ethical Data Use

The final day of the 2016 PASS Business Analytics Conference in San Jose, CA started with a keynote on the ethical collection and use of data by Susan Etlinger, an industry analyst with Altimeter Group, a Prophet Company.  Entitled The Trust Imperative: A Framework for Ethical Data Use, Susan’s keynote addressed public concerns over corporate use of consumer data, from ad targeting and personalization to product development, risk management and beyond.  She was quick to assert that she was not proposing to cease collecting data but rather to consider the ethical aspects of data collection, stewardship, and use at all points in the data life cycle.

Ms. Etlinger opened with a story most of us in the industry know well: the story of the newly-pregnant teenager who made purchases from Target that ended up alerting her parents to her pregnancy through targeted marketing resulting from those purchases. Similarly there was the story from Alexis Madrigal, Contributing Editor at The Atlantic. Alexis and his wife had recently found out they were expecting – so recently in fact that they had only just purchased the pregnancy test and had not told anyone of their news. Yet the very same day they discovered they were pregnant they also began receiving infant-related retail marketing. Surely it had to be targeted marketing initiated by the purchase of the pregnancy test, right? Actually as he dug into the source of the data (remarkably easy to do as explained in the article he wrote on the subject) he came to realize this was a case of supreme coincidence related to buying baby-related items for a niece and nephew four months earlier. Why pre-natal supplies? Studies show that if you capture a consumer’s attention as a brand when pregnant they will be customer of your products for years to come.

Pulling data together from multiple disparate sources can have unintended consequences. Both stories highlight how consumer data collection can be used to make intelligent decisions quicker, and conversely, as Etlinger stated in her keynote “machine learning and artificial intelligence help you make bad decisions more efficiently.”

What about the data we unknowingly offer up from our phones? There is a reason that Google and Apple hold the largest repositories of logistical data. Our phones constantly ping our whereabouts out through location services. There was an article in 2015 about how this mass collection of data is being used to calculate wait times in security lines at airports. We all want to know how long we should expect to wait in security so we can determine when we should leave for the airport, right? This is a good thing. However step back a second and remember that your MAC address is also broadcasted as part of this process – this is the unique identifier on your phone that can tie back specifically to you. This is how mass data collection works now. It’s not limited to opting in or giving your email address willingly to obtain a discount. It’s about around-the-clock collection from your pocket to countless targets.

In a recent study 91% of consumers believe they have lost control over their personal information. But it’s not just consumers who are concerned.  According to PriceWaterhouseCoopers 17^th Annual Global CEO Survey, lack of trust in business is “a major concern of CEOs, with half of them identifying this as a real threat to their growth prospects…up sharply from the 37% who cited concerns last year.” To add, quoting Jennifer Glasgow, Chief Privacy Officer at Acxiom, “Just complying with the law is not going to be nearly enough to make consumers comfortable.” This means there is likely a market opportunity for those corporations and retailers who can demonstrate they are willing to be more ethical and secure in collection of consumer data.

Etlinger used this concept as the basis for her exposition on the crux of the keynote: The Information Accountability Foundation’s (IAF) Principles of Ethical Data Use. The IAF Principles of Ethical Data Use posits that there are five conditions that should be a part of all steps in the data life cycle.

According to The IAF Principles of Ethical Data, data usage should be:

From collection and processing of data and on to analyzing and drawing intelligent decisions from the collected data Etlinger stated that each of these principles should be adhered to. Of course it doesn’t stop there. Storage and security; through to governance, usage and communication resulting from collected data we as data professionals and consumers should be striving to follow the borders defined by these principles in action.

Speaking of borders, Etlinger brought the now defunct Border’s book chain into the discussion with regard to the principle of sustainability.  When Border’s went into bankruptcy Barnes & Noble made an offer to purchase their complete set of customer data. At no point were customers approached concerning whether they wanted their collected information shared with a new corporation – they were not offered an opt-in to Barnes & Noble’s access to their records through transferability.  Does the data we collect ever truly get sunsetted or does it live on?

“By knowing where the borders are, you can innovate more around them.” – Stefaan Verhulst Co-Founder and Chief Research and Development Officer The Governance Lab (NYU)

“Take the reality of where we are”, Ms. Etlinger concluded, “all the technical challenges and (unstructured data) and then have a way to think about engendering and creating trust. Let’s apply these principles bit by bit as we go through the life cycle. As we collect data are we thinking of the benefit? Are we thinking about whether we’re going to collect it over time? As we’re processing it are we thinking about (these principles)? Are people seeing it that should not? Is there respect around the people gaining access to that data? Are there appropriate controls around access to that data or is the data broadly accessible? Because one of the challenges here is that this isn’t about PII or credit card numbers it’s about taking one insight and another insight and discovering a teenage girl is pregnant. Or taking one insight and another insight and discovering someone is standing alone on a dark street corner or that someone is depressed or suicidal...or making an inference about someone based upon the zip code in which they live.”

Different people are involved in different stages of the data life cycle. If they’re all on the same page regarding principles then discussions around how to collect, store, manage, govern, and derive insights from our data become easier and more meaningful. It can lead to more trust from consumers and a more secure ecosystem for this global pool of data that will live on long after we’re gone.