Over 2019 we met and spoke with Heads of Pricing, Chief Actuaries, Heads of Underwriting, COOs and Heads of IT, often from the same businesses. For the vast majority of these, the conversations were the same. They had initiatives being driven by their boards which was to implement and use Data Science in order to gain greater insights to their business, increase efficiencies and ultimately drive better returns and profit. Simply put by many “We need to do Data Science!!”
Now this task at hand is by any means simple.
Saying “We need to do Data Science” is similar to saying “We need to do Insurance”. The breadth and scope of what could be achieved by effectively implementing Data Science into a business is limitless however it’s also difficult.
The C-suites understand the value of using Data Science and see the benefit it has brought to multiple other industries who effectively use it to generate increased profits and find trends in their the business that they were not aware of, not only from a commercial point of view but also an internal one, ranging from;
employee productivity, employee engagement and wellbeing
But there needs to be strategy and plan. And for some of the clients we met with, even by their own admission there wasn’t one.
What’s the strategy?
The potential applications of Data Science in insurance are numerous. From underwriting and loss prevention, product pricing, claims handling, fraud detection, sales and customer experience to understanding risk appetite and premium leakage, to expense management, subrogation and litigation.
Here in lies one of the core challenges in effectively using and implementing Data Science into a business …. which areas do you pick? What will have the quickest and most meaningful impact to the business?
Many clients we spoke with, have hired a data scientist over the past few years however many of those individuals have left. The common theme mainly being, that the business didn’t know what to do with them. With expectations being greatly different to the reality, which is in part caused by a having a lack of readily available data and/or not having the right technical systems in place to enable the individual to the work in which they know how.
The companies that have had a clear data strategy are the ones who have been most successful.
In general, these successful strategies have started off small with one or two data scientists in a specific business unit which has readily available data.
One of our clients did this by bringing on board a data scientist into the sales and distribution business. The individual quickly showed them where and how they could increase revenues. This approach led to quick backing from the business around the effectiveness of a data scientist in the team. Fast forward 18 months and the original Data Scientist now has a team of 10+ and consults as a centralised function to multiple parts of the business.
This strategy was successful as it had a clear goal from the start, had data for the individual to work with and gave them the chance to show their value to the business. Unfortunately, though for many Data Scientists the strategy hasn’t been in place before they start which makes the task an uphill battle.
The major issue is that many had in 2019, simply put, is that they jumped the gun and the results weren’t there. At the beginning of 2020, from initial meetings we have had, the pressure from 2019 to hire Data Scientist has mellowed which would lead one to think that hiring would be subdued for 2020.
However, we are anticipating it to be far greater than 2019. With the pressure reduced from C-suites it is now giving many managers in the market some time to take a step back and develop a clear strategy and approach for what they want a Data Scientist to do for their business and teams.
In part 2 of this article we will be looking into what the significant factors are that need to be address before implementing a Data Science team aside from a clear strategy.
Follow us to read part two next week…