Event: Analytics for Insurance, Europe 2014

This conference was well attended and lived up to it’s billing of attracting insurers, consultants and suppliers from across Europe. With approximately 200 attendees and representation from over 14 different countries, it was a lively and interesting event with strong audience participation.

Six key themes emerged consistently during the day:

a) try, try & try again;

b) fail fast & fail cheaply;

c) data is an asset, manage it as such;

d) insight is more than just data or analytics;

e) cross-functional collaboration is needed;

f) measure your marketing effectiveness.

To kick us off there was a panel of speakers from leading insurers (Swiss Re, Towergate and Cooperative). Each shared how they had developed their analytics capability and seen value as a result. Some were more positive about the potential of “big data” than others, but once again the most applicable examples were using wider sources of internal data rather than social media or other unstructured external data. There were also a couple of examples throughout the day of analytics centres of excellence being created, although the best organisational fit varied. Some still saw analytics as part of IT, whilst others had established this CoE within Marketing, Actuarial or Underwriting functions. None of the organisations presenting had yet implemented a directorate of customer insight, as pioneered by some of the C-Suite customer insight leaders in the most progressive firms. However, there was good practical advice on there need for test & learn, local business understanding guiding deployment of models not IT and the relative benefits of in-housing or outsourcing your analytics capability (a question which recurred during the day).

Following this, we all had the pleasure of hearing a great presentation from MoneySuperMarket.com. Without spoiling the surprise for anyone who has not heard Orlando before, it concerned the experience Norm Larsen as a persistent inventor. Through the story of his persistence, to create a product needed for early space flight, we focussed on the cultural challenge of achieving great analytics (“it takes more than rocket science to launch a rocket”). This presentation is well worth hearing if you get the chance and landed the points (excuse the pun) of both “try, try & try again” when testing & learning with analytics, as well as “fail fast & fail cheaply” which is very relevant for those establishing innovative analytics or database marketing teams.

Several times during the day there was plenty of time for audience Q&A and I’m glad to say that at this event there was active participation. The kinds of topics the audience were raising included creating business cases for investment in an analytics team, the need for attitudinal understanding alongside behavioural analytics, how to achieve “top table” buy-in and the recruitment challenge faced by many organisations. In fact, although people accepted that graduate recruitment and talent development might be an ideal solution, the need for short term results drove the interest in outsourcing. I warned against this unless very effective knowledge transfer is implemented, consistency of personnel and planning in the time taken to become familiar with business domain and your own data. Much of my concern is of course fuelled by my past experience of outsourcing and offshoring analytics.

The focus then moved on to the use of analytics to enable more intelligent pricing. In a session that was well chaired by Celent, we heard from BGL Group, Storebrand and 1st Central Insurance. Some great geeky fun here for the more numerate. We even got an equation on a slide. Applications included credibility modelling, dynamic price optimisation and management of your street price. Given my past experience of customer insight having a key role to play in ensuring senior leadership understand the customer impact of pricing changes, I was particularly taken by comments from both BGL & 1st Central Insurance. The latter has apparently seen from analysis what I’ve experienced in the past, that unfettered optimisation will punish your most loyal/dependent customers and have decided to take an ethical stance of renewal premium rates being in line with new business. This is a huge commercial challenge for large insurers, with large back books, but given growing customer and press disquiet with the pricing differential that can emerge over years, it needs addressing. I was glad to hear other organisations using customer insight to bring to life the characteristics of customers impacted most by statistically optimal pricing.

I was also struck during the day by the level of academic background in a number of the speakers. Beyond the hype of the data scientist role, there does appear to be a real growth in analytics and customer insight (or customer science) leaders having PhDs or coming from backgrounds in academia. This certainly has benefits in the level of statistical understanding expressed by a number of speakers and it was good to benefit from some of their teaching. However, I believe there is also an inherent risk as well. A risk that predictive analytics become more theoretical and focussed on optimal techniques and thus more removed from real world customers and effectively using analytics with research and the learning of front-line colleagues in customer services etc.

Given that concern, it was reassuring in the afternoon to focus on customer analytics. I shared the stage with Christina from AIG and Marion from If Insurance (Norway). The audience appeared to respond well to my presentation and I will soon share my slides via SlideShare for those who are interested. Christina and Marion also did a good job of highlighting the range of application areas for customer analytics, including media mix effectiveness measurement and optimising your multi-channel database marketing through test and learn. The questions we received in a follow-on panel session again revealed an audience with concerns about recruiting, outsourcing and marketing effectiveness measurement.

So, another useful event to attend, and one that will hopefully help me further shape the content on this blog to address the questions/concerns of today’s customer insight leaders. If you attended this event then please share your comments below, or just let us know which of the topics raised in this post you would like to see covered in more depth.