Citizen analytics and the American heartland
- Posted on November 20, 2015
My dad used to love to work on cars. More precisely, he found great satisfaction in making things that were broken, work. He wasn’t one of those guys who had to turn an oil change into a NASA project complete with mil-spec tooling, nor was he one of those guys that figured the engine would surely let him know if, and when, it ran low. His was an American heartland philosophy that focused on core functionality and keeping the machinery operating at its full potential; everything else was either chrome plating or neglect.
I often hear insurance clients talk about not being ready for advanced analytics or conversely their ‘need’ to have an enterprise scale, state-of-the-art analytics infrastructure that permeates all divisions and all levels of their organization. When I ask ‘why not’ to the former group, and ‘why,’ to the latter group, the responses I get usually turn emphatically vague. There is a place for precision, high-end analytics in an organization – usually localized, but today there is also a very useful place for everyday analytics available to a broad segment of users.
We’re in the early days of what some call the ‘citizen data scientist’ movement. Now is the time to experiment and challenge analytic techniques to bring insight and value to specific business problems. We’ll figure things out quickly enough, but to get there we need to avoid both chrome plating and neglect. We need systems that are flexible, have the ability to cost effectively iterate through ideas and approaches, and can scale when winners are identified.
I had the privilege recently of working on an insurance analytics offering grounded in these ideals (see video below). We released it in October and have been getting a lot of interest from carriers who want to apply more, and better analytics to a broader segment of their operations. The offering is aimed at bringing citizen analytics to everyday activities so the core of the business can operate at its full potential.