Personalising customer interactions: the power of advanced analytics in building customer loyalty and satisfaction

Insurers are increasingly applying data analytics to elicit deep insights into customer demands and behaviours. Yet our interactions with them remain largely transactional, leaning towards matter-of-fact updates that urge them to make a purchasing or payment decision.

For example, if analytics shows that customers who have signed on to automatic premium payments have lower lapse rates, insurers can develop a proactive marketing campaign to encourage a shift to that payment method.

Campaigns like these can be effective. But they’re not personalised. Without factoring in customers’ individual needs and context, we risk annoying and alienating them with a mistimed call or e-mail. And if we only get in touch to discuss purchases and renewals, it’s hard to convince customers that we care about them more than their wallets. 

There are well-established links between customer loyalty and insurers’ ability to deliver non-financial value. Consumer loyalty and advocacy increases in every country surveyed when insurers deliver a strong performance on higher-order elements such as reducing anxiety and providing hope.¹  This argues for our industry to reconsider our approach to analytics as they are used to support customer interactions. 

Truly personalised interactions require an understanding of what each individual customer wants or needs from their insurers, and a fitting response in return. While proactive engagements are a good place to start, their effectiveness diminishes over time partly due to the saturation effect. Using the premium payment example - as more customers switch to automatic premium payments, the model gradually moves down the list to target customers with a lower likelihood of responding.

In contrast, incorporating proactive campaigns with personalised activities and deploying at the optimal time can create a more enduring impact. This requires us to know who we are talking to, their key preferences, the most relevant touchpoint for them personally, the channel and time of day they are most willing to engage, and the content they’ll find useful. 

While advanced analytics is commonly used for customer segmentation, we rarely see it used to discover the most opportune moments or methods for customer engagement.  Yet we have the tools to do this today, and the payoff for taking customer engagement to the next level can be significant. 

At Swiss Re, we have been using our Impact+ Retention Optimiser to help insurers optimise their customer retention strategies, and the results have been highly encouraging. We see traditional machine learning models for customer lapse segmentation improve Return on Investment (ROI) by 6 times. More advanced approaches that analyse each customer’s attitudes and behaviours to understand ‘when’ and 'how' improves the ROI to 18 times. If we also use analytics to personalise the communication content, then we can see a further 50% improvement in ROI. 

Applying analytics in this way helps us quantify the ROI of personalised customer engagement. This can be achieved and measured via A/B testing of customer retention or cross selling, reduction in service delivery costs and avoided cost by pre-empting customers from interacting in more expensive ways.

Personalising customer engagement also requires us to rethink the customer journey. While traditional approach tends to see it in clear-cut phases – onboarding, policy renewal and claims – advanced analytics allows an enhanced and more comprehensive understanding of the customer journey.

Returning to our earlier example of a blanket campaign to encourage automatic premium payment. When we extend our models to recognise how customers engage with our products, their premium payment behaviours, indications of shopping around, and receptiveness to insurance advisors' recommendations – we realise that these additional insights can help improve the campaign outcome significantly.

An engaged customer who pays premiums reliably but not automatically and shows no sign of shopping around, for example, may ignore or find repeated reminders to move to automatic payments irritating. But if the model were to identify signals that the customer is looking for alternatives, a proactive phone call to acknowledge their loyalty and learn more about what has changed for them may just make the difference between retention and loss. Applied on a wider scale this demonstrates how analytics can play a critical role in generating business value. 

All said, analytics can only do so much. Even with the best technology in the world, if the content delivered in each interaction falls short of the mark, we'll still fail to realise the potential of our analytics models. 

There are several avenues to develop and evolve the content each customer interacts with, to ensure it’s delivered with maximum relevance via the best channel.

One is to draw on behavioural science to understand the factors that motivate customer interaction and identify areas to reduce friction. This requires a robust use of statistics and tracking numerous A/B tests to show exactly which material and platforms produce the best results. Another area is to use artificial intelligence (AI) to learn and carefully optimise the content each individual customer interacts with.

The application of behavioural science is a substantial and fascinating topic on its own that has been explored in another Swiss Re article Accelerating insurance transformation through behavioural economics.

The strong case for advanced data analytics in enhancing customer engagement is clear. The technology is readily available, and it'll be a missed opportunity if we do not use it to help customers narrow their protection gaps in the uncertain times. On the other hand, while consumers are open to sharing data with insurers,²  the onus is on us to maintain their trust, ensuring that personalisation never tilts into being perceived as invasive, and that in using data the highest ethical and privacy standards are maintained.

In our next article we’ll examine in detail how to use data analytics to deliver personalised customer experiences.