Ethics and Reflection at the Core of Successful Data Science
I often think back to very sound advice I received from my mentor early in my analytics career. As someone that did not have a significant client-facing role at that point in my career, his advice to me as I worked on analytic projects was to make sure I could always explain, justify, and defend each and every decision and recommendation I made as I progressed through the analysis. To put myself in the shoes of the client and fully anticipate and understand their needs and then exceed their expectations. That left a lasting impression on me and conditioned me to always be thoughtful and thorough across all stages of the analytics process: the analysis design, the use of consumer data, the recommended insight-driven business actions, and the measures of success. That approach worked extremely well in a business-centric operating model.
Fast forward to today’s business environment, where customer-centric operating principles rule the day, and it becomes clear that business-centric analytic and data science processes are no longer sufficient. Companies have become obsessed with using consumer data to find a competitive advantage. In fact, Forrester Research explains in their 2020 Predictions: Customer Insights report that 56% of the businesses surveyed will be launching initiatives and appointing “Data Hunters” to identify new sources of data. Personally, as a consumer, I find that a bit disturbing! This increased emphasis on data collection requires a new set of analytics and data science operating procedures to ensure this information is not misused or abused.
That great advice that taught me to anticipate my clients’ needs must now be extended to include another key constituent, the CONSUMER! That is, all of us in the analytics community could benefit from adopting work habits and processes that encourage analysts to step into the shoes of the consumer to help inform and govern our data management and data science practices. To pledge to be transparent and act in the best interest of consumers. To be comfortable explaining to consumers how we use their consumer interaction, transaction, and demographic data to generate insights and how those insights influence our business decision-making and actions. Using this approach, we can not only satisfy business needs, but build consumer trust.
The notion of adopting customer-centric business practices is certainly not a new concept. Thought leaders such as Don Peppers and Martha Rogers have been highlighting the benefits of these principles for many years. In fact, the overarching theme in their book Extreme Trust: Honesty as Competitive Advantage is to “treat the customer the way you’d want to be treated if you were the customer.”
Unfortunately, companies have not always followed these principles. Even more alarming is that several recent corporate transgressions have been linked to the inappropriate use of consumer data. In 2018, Facebook enabled Cambridge Analytica to use millions of members’ personal data without their consent for targeted political advertising. In early 2019, YouTube’s recommendation engine was facing major criticism for making it easier for pedophiles to find and share content on young children. Even more recently, Goldman Sachs has been under fire for having blatant gender bias in algorithms used to establish credit limits for Apple Card customers.
So how do we change the underlying practices that are enabling intentional and unintentional misuse of consumer data before state and federal government data protection regulators step in to do it for us? I believe it starts by creating greater awareness of the damaging consequences of poor data stewardship and reckless analytics practices. This should begin in higher education and be reinforced through recursive training programs in the corporate environment.
After spending 25 years in the analytics industry, and now as a full-time educator, I think it is critical that educational institutions commit to the development and integration of student learning objectives that focus on inspiring and empowering students to use ethical and socially responsible data collection and analytic practices. We must:
teach students how safeguards can be implemented that reduce the risk of deploying unintentionally biased predictive and ML algorithms.
explain how cross-functional data governance teams can be created to ensure diverse perspectives are considered when deciding what data is appropriate to be collected, analyzed, and used to drive business decision-making and AI solutions.
illustrate how many traditional approaches to consumer segmentation using gender, ethnicity, and socio-economic status often perpetuate the exclusion of consumers
challenge students to reflect on the appropriate and inappropriate use of consumer data.
These learning objectives should be a core and fundamental element of every data science and business analytics program.
In 2017, The Economist wrote the article, “The world’s most valuable resource is no longer oil, but data.” I posit that we will soon lose access to this incredible resource if we don’t prove to the resource providers -- the consumers -- that we can be trusted to use this data in a responsible and value-added manner.
I welcome your thoughts and comments. Please check out www.mcguirkanalytics.com/blog for additional blog posts on a variety of analytic topics.