Other parts of this series:
How to make advanced analytics a core business competency.
To unlock the true potential of the data at their disposal, it’s not enough for insurers to rethink how data is gathered, managed and used by the business. That’s essential, of course. But it must be complemented by a commitment to build technology and teams with advanced analytics front of mind.
Most insurers’ legacy technology stacks are complex, and expensive to run. Built for MI reporting and slower cycles of analysis, they’re unable to support the power, speed and precision of advanced analytics. So what needs to change?
We identify three priorities. First, make full use of the cloud. Data scientists deliver better answers, faster, when they’re given bursts of thousands of compute units at a time. Cloud services, with their infinite scalability, provide the burst capacity that makes this possible. They also make sound commercial sense, allowing users to pay ‘by the drink’ for the analytics technology needed in support.
Next, ensure data scientists and data engineers have the right tools for the jobs they have to do – from rapid data manipulation, feature engineering and analysis to ETL toolsets that won’t fail overnight. In machine learning, the statistical programming language R is king. The next generation of data scientists will know it well (and they’ll expect to use it).
Third, design analytics for rapid insight to action. Teams enterprise-wide can be given restricted access to virtual analytical records through modern visualisation tools. Analytics apps can provide employees (from claims handlers to call-centre agents) with straightforward data-led workflows from the outputs of machine-learning algorithms. And APIs, calling cloud-based analytics models, can deliver near-real time responses to high-volume requests (eg providing quotes through insurance aggregators).
That’s the technology shopping-list. Where teams are concerned, the overriding priority is ‘think analytics, not MI’. Far from producing regimented outputs, consistently repeated, analytics teams have to be highly creative and continuously on the lookout for new sources of value. Old management structures won’t work for them.
Top-performing analytics teams incorporate data scientists, visualisation engineers and data engineers. They adopt and adapt the principles of agile-lean delivery, focusing on value creation, championing speed and delivering transparent insights within weeks of starting projects. Crucially, they work best in small, self-organising and autonomous teams, unencumbered by legacy working practices. One vital team member is the ‘product owner’. They’ll be responsible for identifying analytics priorities that align to the drivers of business value and the organisation’s strategic goals, as well as representing the end-users of analytics outputs.
Newly-created agile analytics teams need to plan meticulously and start small, working on priorities that can be delivered (into live environments) within a two to four-week sprint window. If that capability isn’t already in place, and it usually isn’t, then it should be the focus for the first one to three sprints.
Critically important, but often overlooked, co-location of all analytics team members should mean everyone working in the same room, at the same time. That will help to minimise delays, duplications and misunderstandings. Highly-organised lock-in sessions allow for time-boxed periods where product owners review working code, encourage the team, and make immediate decisions on the insights being created. And transparent working practices will ensure everyone is immediately aware of progress against sprint objectives.
I hope this series has shown how advanced analytics can’t simply be grafted onto insurers’ existing legacy setups. Traditional approaches and MI/BI-focused datasets get in the way of the near-real time, actionable insights that should be in scope. Only insurers that build new, analytics-specific capabilities across their data, technology and teams will capture the significant competitive edge that’s now there for the taking.