Retention is having a comeback. For years insurers have used traditional approaches for retention that fell well short of potential, as they tend to be reactive, “one-size fits all” and were activated very late in a customer’s decision cycle. But new capabilities in analytics, raw computing power and behavioral science have created new ways to dramatically improve retention rates.
Driving retention uplift
These new retention strategies could not have come at a better time. In the face of continued commoditization of insurance—with sustained high advertising spend, lower barriers to switch, and changing customer demographics and attitudes that show millennials as the most active switchers1—the US auto insurance industry has witnessed a steady decline in retention over the last decade.
The good news for insurers is that materially significant new capabilities now allow them to proactively engage their customers using new models and approaches that drive a step change uplift in retention (see Figure 1) through:
- A better understanding of behavioral science and the impact it can have on how, when and to whom retention strategies are applied.
- New and more available datasets paired with advanced analytical models that enable targeted, proactive and context-relevant actions at the point of shopping of an “at risk” customer. For instance, there is publicly available information surrounding online browsing data, mobile app usage, card spend data, lifestyle and life events data which, combined with carrier data (for example customer, claims and channel interactions), provides a rich base for applying the analytical models.
A meaningful shift
A critical path to customer retention relies on generating highly accurate churn models. We apply an extensive, proprietary 360-degree customer data set which includes external digital, financial and locational information on 250 million consumers in the US that can be combined with a carrier’s internal customer data to generate highly accurate churn models. Specific machine learning algorithms can then be deployed against these data sets to power prescriptive analytics assets to develop proactive, context-specific retention tactics. Finally, combining these analytical insights with human-centric design drives innovative and personalized strategies that are much more effective at retaining the customer. And customers are ripe for this more personalized attention, with 81 percent of customers who switched saying their provider could have done something to keep them,2 and nearly half (49 percent) indicating they appreciate when a company provides an offer or other service specifically personalized to them.3
Specifically, we implement these capabilities through a structured, four-step approach:
- Predictive analytics: The most important element in a retention program is identifying with a high degree of certainty which customers are shopping. Using machine learning we build highly accurate models that predict which specific customers are beginning their shopping process and which are most likely to attrite, allowing an insurer to understand exactly where to place proactive retention efforts. With this analysis, a carrier can be much more targeted and spend much more effectively on tactics that will influence a customer’s behavior and decision to stop shopping.
- Experience design: Customers start their shopping process after being “triggered” by one of four primary factors: experiencing a rate increase, enduring a negative customer experience, simply being exposed to competitive marketing, or having a circumstance that requires an update to their policy (e.g. new car, adding a driver, etc.). Based on this, we design human-centric, behavioral science-based retention tactics, specific to each shopping trigger that can be deployed to each at-risk customer.
- Prescriptive analytics: With a set of treatments developed, the next step is to analytically test them both individually and in various combinations. A/B testing is used to determine which sets of treatments generate the greatest uplift in retention.
- Adoption and sustainability: Once the models and treatments are developed, the process will need to be operationalized on a weekly or even daily basis to identify customers that are at risk and begin implementing the system of tactics that will convince them to change their shopping process.
These strategies can help drive substantial revenue potential: in our experience, applying these new capabilities with insurers has resulted in 3 to 7 percentage point increases in retention among the at-risk customer base (customers that begin shopping for insurance). To put that in perspective, an insurer with $10B in annual premium growing at market rate could add over 6 percent in incremental annual premium within five years by implementing these strategies. Cumulatively, this would add $2.7B to the insurer’s premiums over the five-year period (see Figure 2).
Further, given that retaining a customer is typically less costly than acquiring a new customer, a retention approach similar to this would have a disproportionately high impact on the insurer’s margin—presenting an attractive formula in an industry where companies struggle to gain share without sacrificing profitability.
In the zero sum game of insurance, those that ignore making step change improvements to retention risk losing share to the competitors that do.
1 Accenture Strategy Global Consumer Pulse Research, 2017
3 Accenture Strategy Global Consumer Pulse Research, 2018