One of the biggest issues we face with our analytics projects is the challenge of change management. I have seen this throughout my career, change management issues typically arise when there is a desire to develop the business process, normally when we are implementing a new technology platform.
But why is this so prevalent in disability claims departments? And why are claim managers so resistant to change?
1. Unique breed – I was discussing this topic with the VP of Operations from one of our Canadian clients and she was very clear about the issue. They recruit a certain profile of personality for their claim management positions. The individuals need to be questioning, judgmental, and independent thinkers. These personality traits are brilliant for claim managers, but they make implementing change challenging.
After some thirty plus projects, I am starting to understand some of the influencing factors on this strange phenomenon.
2. ‘Stop giving me numbers!’ – Claim managers are very busy people, the better work they do, the more claims they get to manage, and then they are expected to mentor other less experienced individuals. So, don’t give them anything that makes their life more complicated.
3. Half-baked ideas – Unfortunately the Claims Dept. is not ‘top of the tree’ when it comes to investment, and experienced claim managers have seen too many half-baked ideas get implemented in the claims operation. On this basis, claim managers may be just a bit tired of consultants who think they are smart, messing around with their tried and tested process!
So, my data science friends, now go in there and explain, with lots of long words and complicated probabilities, how your latest predictive model is going to change their lives!
How about we try something different? Instead of starting with the data and looking for patterns and performing multivariant analysis… Why don’t we start with the claim managers and the claims process, and gain an understanding for what challenges are faced in the day-to-day management of claims.
Instead of delivering the answer, we first need to define the problem.
Then we can direct our data analysis toward helping to improve the process, and we will now have buy-in from our claim managers. Provide actionable information, not numbers and probabilities!
To a certain extent, the problem is exacerbated by the incredibly sophisticated data modeling tools which are on the market today. You can shovel the data in one end and get a choice of models out of the other end with varying methods of ‘pruning’ for model optimization. It’s too easy to build a model without a real understanding of what it’s doing.
To Claim Managers – Be as cynical and questioning as you can to shake-down the data scientists.
To Data Scientists – Tread really carefully, understand the business process, focus on issues raised by the claim managers, get buy-in to what you are doing, don’t get carried away by your ‘wiz-bang’ modeling software.
To Operations Managers – Try to keep the peace… and then enjoy the amazing results!
I hope you find this topic thought provoking.