Yesterday I began an exploration on creating better data science in the compliance function based upon a Harvard Business Review (HBR) article, entitled “Data Science and the Art of Persuasion”, by Scott Berinato. Yesterday, I introduced the problem and how to build a set of talents to a better integrate a data science operation in your compliance function. Today, I want to consider how to implement those talents into an integrated team by having the right set of team skills.
This is still a tricky area for most legally trained compliance professionals as law schools are way behind the business world in teaching these skills. Yet, not only data analysis but also the presentation of data in a visual format will be a key skill for every Chief Compliance Officer (CCO) and compliance practitioner going forward. In the prior post, I detailed some of the talents your data science operation will need going forward. Recognizing you may not possess those skills yourself, the author suggests creating a ‘kitchen cabinet’ of experts you can call upon to help you going forward.
The author believes, “Thinking of talents as separate from people will help companies address the last-mile problem, because it will free them from trying to find the person who can both do data science and communicate it.” Some of the talents previously outlined naturally can go together, such as design and storytelling or data wrangling and data analysis. You may even be lucky enough to find that they exist in one person. The bottom line is that finding some “people who have superior design skills will free data scientists to focus on their strengths. It will also open the door to people who might previously have been ignored. An average coder who also has good design skills, for example, might be very useful.”
The communication is just as critical as the data analytics for if you cannot explain it, you cannot make use of the insights. The author cites Randal Olson, the lead data scientist at Life Epigenetics who previously focused solely on how well someone did the technical part of data science. “I know, when I started, I had zero appreciation for the communication part of it. I think that’s common.” Now, in some cases, he has changed the hiring process by occasionally bringing in “a nontechnical person and say to the candidate, ‘Explain this model to this person.’”
If it is not obvious by now, the skills used by lawyers and compliance professionals tend to be very different than data scientist. However this does not end your effort or even your inquiry. A good reminder for every compliance practitioner is that “overcoming culture clashes begins with understanding others’ experiences.” Those presenting the data, most usually the CCO or other compliance professional, are usually not well-versed in statistics or algorithms. Certainly the lawyer training in me would focus on the narrative. As the author notes, the “depth and complexity of data work is hard for designers to reconcile.”
Statistically trained “data scientists, in contrast, value objectivity, statistical rigor, and comprehensiveness; the communication part is not only foreign to them but distracting.” The author quoted one manager of a data science operation at a large tech company. “I was the same way, working in data science for 10 years, but it was eye-opening for me when I had to build a team. I saw that if we just learned a little more about the communication part of it, we could champion so much more for the business.” There are many ways to expose your compliance function team to the value of others’ talents. The compliance professionals should learn some basic statistics, while data scientists learn basic presentation principles. The goal is not to become experts in the counterparts field but to learn enough to appreciate each other.
The key is to present the data in a way that everyone in the audience, typically senior management but it can be down to the front-line business unit level, can understand. This is where the storytelling skills come into play. Your meetings to deliver the results should always include a mix of skill sets. A meeting should, therefore, include someone with the presentation skills of a marketer who can sell the insights and make the presentations. However there should also be subject matter experts who bring data wrangling and analysis talent to meetings. Special sessions at which stakeholders put questions to the full data team can also be included.
Cross mentoring can also be useful. The author states, “People who express interest in developing talents that they don’t have but that you need should be encouraged, even if those strengths (design skills, say) are far afield from the ones they already have (data wrangling). Indeed, in my workshops I hear from data scientists who would love to develop their design or storytelling talent but don’t have time to commit to it. Others would love to see that talent added to their teams, but their project management focuses primarily on technical outcomes, not business ones.”
This type of mentoring can also foster greater teamwork around data and data analytics by creating empathy among team members with differing talents. This in turn creates trust, a critical element for effective teamwork. The author quoted, Eric Colson, the Chief Algorithms Officer at Stitch Fix, for a situation where he used storytelling talent to help explain something coming out of data analysis: “I remember doing a presentation on a merchandising problem, where I thought we were approaching it the wrong way. I had to get merchandising and sales to buy in.” Instead of explaining beta-binomial distribution and other statistical concepts to bolster his point of view, he told a story about someone pulling balls from an urn and what happened over time to the number and type of balls in the urn.” It created such a level of trust that he did not have to explain the math behind it all.
With a portfolio of skills ready to go, you can now move forward to use the data in a meaningful way. In the final post, I will conclude by bringing it all together for the CCO.
[View source.]
