Managing insurance customer recruitment budgets to optimise longer term customer value
This case study is important for insurance intermediaries or product providers because it:
- provides an approach that allows you to allocate recruitment budget by channel and customer type to optimise longer term value
- shows how a range of factors looked at in combination can be used to maximise marketing profitability or revenue
- reveals that once you understand how different customer groups perform you can also target inside your recruitment channels
Case Study: Using data science to allocate multi-channel marketing recruitment spend towards longer term customer value for a large life intermediary
This case study has been based on the experience of working with a substantial life insurance broker who has been using multiple channels to recruit customers in the UK. The products are predominantly life protection and the broker is remunerated through commission.
We were asked to look at how best to target customer value as judged by the commission earned, after claw-back, in the first 24 months from recruitment.
As will be familiar to life insurance intermediaries and providers, lapsing early (i.e. in the first 24 months since inception) was seen as one of the main causes of loss of customer value.
Lapsing early was found to be more prominent among people recruited via outbound direct marketing channels than through more self-driven channels such as inbound to web. It was also correlated with the sum assured and the monthly premium. However outbound recruiting is cheaper, and higher premiums bring more commission.
The relationship between lapsing early and age was non-linear, with both younger and older people lapsing more than those in the middle.
And as well as different recruitment channels having different costs per acquisition, they also attract different age groups and sums assured.
How did we set about enabling our client to choose the right recruitment tactics to deliver the highest longer-term return on recruitment marketing spend, with so many factors at work?
We started by developing a customer group contribution table that contained predicted longer-term revenues for each micro-cell.
1. We discovered that lapses in the first six months are very reliable predictors of the propensity to lapse within 24 months for every customer group. (see Fig. 1); this meant that we did not have to wait for 24 months’ worth of history before making a prediction of claw-back relevant lapse rates.
Fig. 1 Plots the cumulative occurrence of lapses up to 24 months
2. After examining many possible factors, and eliminating those that were less significant, we defined the customer groups to model by a combination of age range, sum assured, and recruitment channel.
3. We then used a combination of known lapse experience with predicted to build the overall expectation not only of overall lapse rates but also when lapses would occur.
4. The average net contribution for each customer in one of the cells below (see Fig. 2) was computed as predicted commission income for 24 months after claw-back and after recruitment costs
5. Therefore, for each customer group, we could predict its net customer value; we also knew the historic number of policies being sold in each group and hence we could calculate their overall predicted future value.
Fig. 2 This chart shows the net contribution by customer in each group over 24 months (the actual numbers have been changed into colour bands where the highest contributors are green, lowest positive contributors yellow, and loss makers red)
6. For each channel we could see where the profitable and unprofitable customers came from and hence:
– how to target recruitment within the channel or drop the channel altogether
– the volumes historically being obtained in the profitable parts of the channel
7. By then examining the extent to which spend could be increased in each of the more profitable sectors, and taking account of saturation effects inherent in increasing spend in any sector, we could allocate spend across and within channels to optimise longer term value.
8. Return on the investment in data analytics?In this example c. one quarter of the recruitment budget was reallocated from unprofitable to profitable activities. Prior to the reallocation of the marketing budget, the overall return was£X, and after reallocation the overall position was £Y, with the analysis cost of £Z. The return per annum from the analysis was (£Y-£X)/£Z. In our experience the ROI on the analysis cost is often greater than x25 in one year.
First, we strongly recommend a multivariate approach when looking at the effectiveness of different marketing tactics; that is, allow for the full effect of all factors in combination on future net customer value to be taken into account. In this way the numerical statistical analysis most closely reflects the real world.
In this case, the focus for this client’s recruitment turned away from broad brush channels like daytime TV towards more direct and controllable channels like Facebook and direct mail. This switch made a very substantial difference to overall returns.To get the benefits,businesses need to be prepared to implement significant changes to current budget allocation for customer recruitment.
We now update the predictive models, and customer group contribution table, on a quarterly basis to take account of new channels being tested and changing market conditions.Thus,the analysis work runs in parallel with the budget re-allocation and keeps track of its impact.
In the words of our client:
‘We have worked with Berry Thompson successfully for 2 years. In that time their team have played an instrumental role in converting reams of our data into deep pockets of useful information. Part of their success is their ability to listen, adapt, and convey results simply, from which solutions can be sought. Unlike many consulting groups, they are customer focused and solutions driven. They have added measurable value to our organisation’.