Share this content

Big Data and analytics: 10 lessons from the insurance sector

15th Apr 2016
Share this content

Falling customer satisfaction scores, margin erosion, increased customer churn and a highly competitive market. All this, plus a consumer champion regulator who continues to raise the bar and demonstrate their ability to fine.  I know it's hard to feel sorry for insurers, but it's a tough market in which to prosper these days.

Those who are taking market share, amidst the carnage, often demonstrate a focus on customer insight and innovation. Their business models differ, including not just disruptive start-ups, but also ancient traditional 'life' insurers. What they appear to have in common is an ability to derive real value from their data and analytics.

Drawing on a recent conference which covered how data, analytics, research and database marketing can make a difference (when combined with the right softer skills), I'd like to share with you 10 voices. Each comes from a different part of the insurance sector, and each had their own lessons learnt to share:


Kirill Pankratov, shared the importance of a good relationship with IT. Plus evidence of improved value across the customer journey from applying analytics (including understanding them for marketing, smarter pricing from risk modelling for acquisition and renewal and improve claims decisioning. The key lesson here seemed to be about thinking across the customer journey. How might you apply more data and analytics to improve acquisition, education, cross-sell and what Kirill called “recovering the relationship”? 

Swiss Re

Daniel Ryan, focussed on the variety of R&D tests and applications they are implementing. From A/B testing of comms for behavioural biases (as this blog has advocated), to application of IBM Watson machine learning to reinsurance problems. Some of the most interesting applications were use of new datasets created by customers (inc. and the opportunity for reciprocity in data sharing and services provided by wearables and other IoT devices. The key lesson I heard was the need to build trust with customers and recognise them as owners of their own data, so instead of big data by stealth, create apps that provide services they would value in return for data sharing. What could you offer your customers as a service or insight into their lives in response to data you’d value as an insurer?

Having an opportunity to present during the main conference as well, I shared a short overview of our training on “Consultancy Skills for Analysts”. It was encouraging to see again the interest from many of the data and analytics leaders there in developing their teams softer skills (like questioning, stakeholder management, planning, facilitation, communication and influencing). The 9 step model we use for structuring this training still appears to have wide application. Key lesson? Do you need more data/technology, or would you get more return by investing in developing your analysts to have more impact in your business?


Paola Scarabotto, gave us an Italian perspective. She echoed previous comments on the importance of context, both for analysis and for communication and services for your customers. Using Telematics data, combined with other public and social data, has enabled the development of services to help their customers when needs arise (collision for example). The main lesson was the importance of context, or as Groupama consider it, being a “Proximity Company” for their customers. A trusted partner to help when things go wrong. How could your data enable closer help in context for your customers when they need it, but without intrusion at other times?


Steve Jackson, came from a different perspective, as an ex-policeman with responsibility for detecting financial crime. Here the need for data sharing between insurers was a key theme, even after years of efforts from CUE, MIAFRA etc. The breakthrough for them appears to have e been real-time analytics during a claims call. Using questioning and validation using internal and big data to create fraud alerts where needed or confirm those who can be paid quickly. A key lesson was that sometimes realtime really does matter, reducing customers down to segments or batch scored propensity scores is not as powerful as the data being able to provide active listening and prompting to a customer conversation. Do you have customer or business needs that would benefit from realtime insight?

Standard Life

Howard Barber, shared the journey of this leading pension provider. Entitled “The Art of Good Conversation”, Howard explained their cultural journey to go from ‘Customer First’ as a vision, to analytics and customer insights really being embedded in decision-making across the organisation. A key lesson I took from this, as well as recognising many of the challenges, was the theme that every interaction matters. This has been true for them both with regard to customers (moving to A/B testing as norm and better coordination of messaging journey) and in embedding use of insight into leadership behaviour (moving to pull not push analytics anddemonstrating ROI through stronger marketing measurement). How are you influencing your senior leaders to see the value of insight-led decision-making?


Reza Khorshidi, comes from a more academic background and so takes an R&D approach within AIG’s Science team. Good to hear his call for the importance of storytelling, plus the importance of visualisation in communicating analytics results. There were interesting points on Simpon’s Paradox, optimal use of your data scientists and the potential for Insurance to be the next frontier for FinTech. But for me the key lesson was very similar to that shared by Daniel, the opportunity of Me2B economy (buying apps to help me in my life through the ‘payment’ of sharing data I generate in my life). Are there new commercial models for your business whose currency could be customers’ own self-generated data?


Kiran offered a somewhat contrarian approach to using data to gain a commercial advantage. Some was clearly beneficial, using big data to improve ID and verification of customers and to identify those attempting to “game” quote engines by altering the data they provide (as apparently condoned by Martin Lewis). However, the key lesson for me was to be careful how you position your data and analytics plans. Kiran presented their potential use of social media and other big data to help with profiling and risk rating customers as “weaponising Big Data”. Given the planned FCA thematic review on just this topic, that was at the least ballsy. Are you careful in the language you choose to envision your business with the potential of big data and analytics?


Barry Hawkins, provided the wisdom of a lifetime spent working within the Insurance industry. There were a number of tips and wise warnings against being sold on vision or shiny new technology, rather than the analytics which actually matters commercially. I also found myself agreeing with Barry on the need to bring data and analytics understanding ‘in house’. Like Howard, there were also war stories from the work needed to move from a product-centric to customer-centric business. A good question proposed was: What is an acceptable intrusion into people’s personal data to provide underwriting advantage? But for me the key lesson was to develop a workable Customer Lifetime Value score and use that to prioritise your other models. Are you making sufficient use of value-based-segmentations (whether for underwriting or marketing purposes)?


Colin Lethbridge, reminded us of the role such work can also play in Solvency II projects. Although potentially a dry subject that people avoid (a bit like data regulation), there is an opportunity here. The key lesson was to spot where something that needs to be built for Solvency III (and thus is more likely to get budget), can also have wider business benefits. Are you joining up with your Solvency II team to see how your intended data strategies could be rationalised, to avoid duplication and reduce time and money?

1st Central

Dilip Singh, was mainly hilarious and should be booked as an insurance industry stand-up comic (perhaps in a double act with Barry, his mentor). To help us end with a more human perspective, he shared his own career journey as an analytics leader and the benefits of working for a smaller private insurer like 1st Central with a start-up mentality. If you get a chance to speak with Dilip, do ask him about Dilip’s 5 laws, but for me the key lesson was maintaining an inquisitive attitude to internal data. He will wander around in his business, see what people are doing and always be asking “can I use that data?” Do you still have a restless curiosity about data opportunities within your business? (To date the biggest returns shown for “Big Data” have come from previously unused internal data).

I hope that helps those leading data, insight or analytics teams within insurance businesses. A number of those lessons also have much wider application across FS and other sectors.

If you have been to any useful conferences recently, what did you learn to help your leadership?

Replies (0)

Please login or register to join the discussion.

There are currently no replies, be the first to post a reply.