Management Lessons

Building an analytics team in the insurance industry, an industry that has been driven with data long before “data-driven” became a popular term comes with its unique challenges. To successfully build an analytics team that will add value in a company that has been using data as part of their key operations for the past 180 years, requires some innovation without being disruptive. I have learnt a lot so far, but here are 3 things I will like to highlight:

People always come first

As a manager, the performance of the team determines your performance. In other words, your team members make you look good. So it is important to invest a very important part of your time and thoughts into the well-being and growth of the team members.

Transparent communication and over-communication can not be overstated especially during this pandemic period when everyone is working from home and facing diverse challenges. A key part of communication is listening. The team members need to know you will listen to them even though you may not have any solution at the moment (which in many cases, I don’t). Working together to come up with a solution helps build trust.

Similarly, setting a growth plan for each team member helps to motivate and inspire. In my case, the team is made up of individuals from different backgrounds on different career trajectories. They all have different skillset and areas they want to learn and grow in. I believe, the best way to learn is by doing. So, I constantly have to do that mapping between the analytics work that align with the company’s strategic objectives and the growth of each team member. It is quite a challenging mapping but it is doable.

Be agile about being agile

I borrowed the above statement from a former boss at IBM. The field of analytics is pretty much changing and with different job titles that have vaguely defined responsibilities. Also, the structure of analytics teams vary wildly - some are under engineering teams, some are a unit on their own others are decentralized across all business units. These leads to no industry standard for running analytics teams so you have to come up with the best way that aligns with where analytics sits in the company.

For example, the analytics team might sit in the company in a form of data support capacity where the team works on adhoc requests. In this case, running sprints might not be the best of ideas because you do not have control over what tasks you will be getting or the urgency of the tasks. You have to figure out a way to be agile about being agile. I will advise starting simple and start tweaking. The change needs to be gradual and there should be some creativity in customizing some well-established processes so you can arrive at what works best for your team. Think of it as hyper-parameter tuning in a model.

Measure Everything

Having the trust and support of the executives will go a long way in establishing any analytics team and the language most executive teams understand is based around a quantitative measure. The team needs to add value to the company that affects the bottom line in a measurable way. This goes a long way in maintaining the trust and support from the executives. Also, the value being added needs to grow so you have to be on a path of constant improvement.

Peter Drucker said, “what you can’t measure, you can’t improve”. You need to setup a framework to measure the team’s work so as to know where to improve and the clearly communicate growth.

In conclusion, transitioning from a core analytics developer to analytics management has been a very interesting ride. I must admit that it was difficult (and strange) spending more time with PowerPoint slides and emails than with Python and SQL code. Nevertheless, I am excited by the impact being made by the team and the measurable growth that comes with it. Will keep on writing about my experience. Stay tuned!