by Matt Ariker
Data science is a relationship game. Sure, it’s about numbers and models and analytics, but as data scientists in even the most advanced, digitally-savvy organizations can testify, an inch-thick model library and a shelf full of advanced degrees only gets one so far.
No matter how potentially game-changing the techniques are, if you can’t persuade others to use them, and you don’t have a seat where the thorniest business problems are discussed, then all that specialist expertise doesn’t matter.
Effective relationships aren’t built through the occasional lunch or division-wide meeting, though. They’re built in the course of doing something concrete and when parties are willing to invest some effort. That’s when people become believers. And when folks who have only the most tangential understanding can finally see the power and versatility of analytics and all that your team can bring to the table, they become advocates.
Few things are more concrete than delivering insights that lead to practical actions and real impact. I’ve been in this business for a while and it’s here where I see many data scientists fumble. They develop incredibly cool models and algorithms, but too often leave them at the door of the business and walk away. They fail to use that moment as an opening to work with the business and help them learn how to use the tools effectively. They don’t guide their colleagues on the processes needed to make best use of the solutions created. And they don’t use their considerable analytical chops on themselves to test, measure, and refine results.
To be heard, data scientists need to solve this last mile influence problem. Here’s how:
- Focus on the all the steps that need to happen after the analytics are created: the processes required to integrate insights into action, the automation to help reduce the burden on teams that are already over-stretched, and the measures of success to evaluate insight efficacy and impact.
- Use those steps as a way to invite business teams into the conversation. Because you’re there as an enabler to help them be successful and resolve their pain points, they’re more likely to see you as a champion—and less likely to resist change.
- Track the outcomes, report the results, and use them to create a virtuous cycle. Develop an executive dashboard that showcases both trends and performance against plan. Use test and control methods to chart revenue impact and attribution. Monitor qualitative factors, too, such as adoption and the number of analytics projects initiated as a result of various management and program meetings.
Putting this all into play isn’t as hard as it seems. For example, at one large telecoms company, the lead data scientist was clearly brilliant and accomplished. Yet he struggled with building strong business relationships, often focusing on the mathematical excellence of his teams output versus what the sales teams actually wanted. So his incredible models were falling flat.
To address this issue, he develop his personal marketing plan. He focused on working with key constituents to identify their concerns with his models, which were getting in the way of implementation. He then made a number of changes, many of which were simple, to show he was listening. For example, changing the UX design to use color codes to show which analytic insights to focus on helped teams understand what to do. The process also made the teams co-owners of the solution so they become more invested themselves in making the analytics work.
By focusing on the last mile of influence, data scientists can not just show their value but see it happen. When data scientists build up their understanding of the business, and forge the deep, collaborative relationships with sales teams, they can embed analytics and an analytics culture deep into the heart of the organization.
Matt Ariker is an analytics partner in McKinsey’s San Francisco office.