Work — September 19, 2019

Don’t Let Big Data Get in the Way of Experimentation

More data isn't always the answer

by Tim Angiolillo

Data Insurance

Big Data Image

Cake & Arrow works almost exclusively in the insurance industry as a customer experience partner for carriers, brokers, insurtechs. I attend most of the major insurtech like this week’s ITC Connect Conference and the C&A team regularly attends various insurance meetups – so when it comes to insurance and trends, we see a lot of conversation across the industry on what needs to change, where we collectively need to go, and what must evolve with the customer experience as we get there.

You can’t observe or participate in this conversation without encountering a lot of talk about data. Customer data. Forward looking conversations on what customer data you can collect, and how the increased ability to gather, disseminate, and act upon it are going to change our industry.

Thanks to smart sensors, beacons, and a plethora of new technology, the volume of customer data has exploded and continues to. Recall the shocking statistic that 90% of all data was created in the past two years – nothing has curtailed or decelerated our appetite for and ability-to gather massive amounts of data from a widening set of sensors and devices. So, insurance isn’t wrong for heavily focusing on customer data – there is an obvious well of opportunity there. But…

I’ve noticed a scary trend. A fear to pilot. A fear to try. A fear to learn lessons from customers and validate experiences without large datasets to support your conclusions. There is a perception that decisions made without large data-sets just aren’t as strong. Two things about that perception give me pause:

1: Large quantities of data need a large investment in analysis to make sense of it all

2: Poor understanding of large data sets leads to poor decision making

Within insurance, a majority of large IT initiatives that promised to do this for customer data aren’t going as planned. If data doesn’t provide flawless decision making and your ability to gather data in the first place is compromised, you have an even greater incentive to seek alternative methods methods for quickly and conclusively validating new ideas with smaller data sets.

“Great, what should I do?”

Reframe the thought and focus on what data is supposed to be providing you – a confident perspective on a given phenomena. An informed prediction of what might happen in the future based on what has happened previously.

As a human-centered customer experience agency, research and testing are fundamental to all of our engagements. We’ve conducted many studies and employed a wide range of research methodologies at varying levels of investment over the years – everything from large, global research events to the humble >10 question survey.

This approach requires a mindset shift; research and data are going to help you develop a validated hypothesis to bring to market, versus data validating the successful elements of an experience in-market, after launch.

I like this quick, idea-generating approach for a few reasons. First, product teams are able to leverage a smaller group of users to gain insights. Actionable insights can be derived from a survey with as few as a hundred people and qualitative insights can be derived talking to a handful of people. On the other hand, when you’re relying on behavioral data from a system in-market, peoples’ behaviors are limited to what you have provided them. Without the ability to follow up on and understand they “why” behind the data or user behavior, you will not have a full picture into areas of pain and potential areas of opportunity in the experience.

The way this functions in practice is you perform user research events with in-progress prototypes / proofs of concept in advance of launching an experience, contrasting that with using analytics to understand user behavior from a live experience.

As you prepare to spend a few days with the insurtech community in Las Vegas, think about your organization’s goals, projects and where you’d like to be in both the short and long term. Expensive, labor intensive methods (read: Big Data) may not drive the kind of outcomes you’re looking to achieve in the here and now. An openness to small scale research and getting to market fast can drive the kind of innovation that keeps you ahead of the curve.

We love talking through this, I’ll be on-site at ITC along with our CEO Josh Levine and Product Lead Julia Molloy – would love to sync up and see how we can help put this into practice.

(Articles only appear in the frontend.)