I live in New England, not an area typically known for its propensity for natural disasters. But this year we've seen record snowfalls, tornadoes, floods, an earthquake, a hurricane, and finally, a Nor'easter (the winter version of the tropical storm) that knocked out power to more than three million customers across the region and decimated neighborhoods of old growth trees.
More than ever, I began to appreciate the importance of communication: Providing information (not data) at the right time in the right channel so I can make good decisions. During the days we were without power, I found it difficult to get information: Which roads were closed? Where could I get supplies? How long would we be without power? The primary channel for information turned out to be television. Since the power and the cable lines were down, it didn't seem like the best vehicle for information. There were no coordinated websites between the state and local governments; and while the utility company provided an outage map on the web, all it told was the percentage of households without power, not when we would have power. That's not information I can use to make a decision.
In my circumstance, I'm pulling pieces of data from all over the place and trying to make sense of them; nobody's telling me the whole story. Who even knows how much of that information can be trusted? The point is that this is a typical problem within many organizations. The internal information providers don't spend enough time understanding what decisions their internal and external customers need to make, and communicating that information to them in a way that's actionable. Data gets thrown all over the place and it's up to the end user to craft a story out of it.
Those of us who work in the information provider business as data or as statistical analysts are often abstracted from the end-customer. This distance often creates a void in our ability to communicate relevant information. I spent a number of years working in large insurance companies, and many roles that I had did not require any interaction with agents or insureds or claimants. In fact, most of the time it wouldn't have mattered if we were selling widgets or insurance policies -- I was that far away from our end-customer. It wasn't until I began to understand the entire information flow across the business that I began to be able to truly craft and communicate a meaningful story to different layers of information consumers.
Many people are also familiar with the analyst's variant of the Pareto principle: We spend 80 percent of our time on data preparation and 20 percent of our time on analysis. Little, if any, time is allotted to communication and presentation of results. How many good analytic projects have failed because the results couldn't be communicated in a way that was easily explainable? For example, insurers can create all of the fancy statistical models they want to optimize pricing and underwriting, but if the agents and brokers that sell your products don't understand it and can't explain it to their customers on why their insurance rates went up or down, you're sunk! How many other great statistical problems have been solved, only to have the results buried because business stakeholders didn't understand them?
Here are my recommendations for business intelligence and analytic teams to help them to close these gaps:
Ask who, what, why, where, when and how when you're solving a business problem with information. By answering these six questions, you can begin to bridge those gaps in your information flow. Understanding what the purpose of the analysis is, who the audience is, why it's important to the business. These questions are essential to the qualitative aspects of the analysis - how the insight is presented, communicated and shared. Then begin to address the quantitative aspects of the data -- where is the trend occurring, how much or how many times do we see it, when is it happening? Many analysts forget to ask the first three questions: who, what, why.
Feedback is key. Make sure that you're consistently meeting with your information consumers and getting feedback. Understand how they're using and interacting with information and what it's being used for. Gauge their technical and analytic acumen and identify ways to make it easier, not more complex, to answer business questions. One of the primary reasons "spreadmarts" continue to proliferate is that information providers either don't understand or can't address end user needs, so end users cobble together information on their own. As an information provider, I would want to know what we're missing.
Become internally and externally customer-centric. On a recent analytic process improvement initiative I worked on with an insurer, the analytic team's customer was the marketing department. Before the team began any additional work on their process, I told them to interview or survey their marketing customers to ensure that their improvement project aligned with their internal customer's needs. Is the marketing department's number one goal flexibility or speed or quality? What are the tradeoffs in achieving each of those goals? Would the answer to those questions change how you "optimize" your process? Even more broadly, how do these targets align with broader business goals around revenue growth or profitability? Becoming customer-centric is the key is to ensure that you're solving the right problem for the right reason for the right customer.
Spend a "day in the life" with a customer. Some of the most rewarding and insightful time I've had with my customers is spending a day in their shoes. I've shadowed claim adjusters, spent time listening in on customer calls in the call center, been in on focus groups with wholesalers. Spend time with your customers in finance, underwriting, IT, operations, wherever, and find out what they do! Not only will you strengthen the relationships with those teams, but you'll come away with a deeper understanding of their business process.