This submission is all about how creating a bit of friction into a digital journey can actually be an effective nudge to enable people to slow down and engage their System 2 thinking. By adding an additional step to a debit-order reversal process, we were able to prevent customers from making a decision that would have negatively affected their future financial outcomes. 

Real-world applied behavioural science is not like the idyllic research papers we so often read. The environments are complex, the data is imperfect, execution risk is high and it is often difficult to control for external factors. You need to be able to figure out how to make behavioural science work to the best of its ability within those environments, as opposed to trying to only engage in areas where everything is a perfect fit. Don't be scared about a complex and messy environment, be creative with how you design, iterate and measure as some of the most impactful interventions will likely occur in imperfect environments. 

Be deliberate about writing up case studies around interventions that make a difference in your world. In my team, I ensured that all of us had to write at least two case studies per year and embedded that as one of our core KPIs. Not only did it help us build an impressive intervention library, but it also increased the diligence around reporting and measurement for all of our work as everything had the potential to be published and shared. Carving out dedicated time to write these case studies and enter the GAABS Awards is not only valuable for your career and your business unit, but it has great spillover effects on the rest of your work. 

A lot of the work we were focused on at the time was centred around insurance and asset management. Often, our behavioural science team was pulled into projects whereby we were working with a relatively short-term goal or battling with some data quality issues which made both prediction and experimentation difficult. This entry was different in the sense that we were able to run a really sound and robust experimental process with an insurance retentions team who clearly saw the value of the behavioural science approach in a data-rich environment that allowed us to be very precise in both set-up and execution. 

Our submission detailed a simple, but extremely effective messaging experiment to increase the first insurance premium payment for funeral insurance customers. The experiment was very successful and we saw the increase at scale as well. We were able to articulate what the increase led to in terms of retained policies and gross written premiums saved for the insurer. The numbers were really impressive to the extent that the intervention managed to retain nearly as many customers as an eight person call centre over the period of the year. We were able to extend the approach and methodology to multiple products and various points of the customer lifecycle. 

From the work, we were able to create an experimental protocol for messaging experiments that was highly scalable and very generalisable across contexts and geographies. We created a process around how to benchmark data, run initial analytics on customer profiles and design experiments with principles that had been tested and validated as persuasive over time. We were able to apply this methodology beyond South Africa. For example, we applied the methodology to digital lending in Nigeria and saw triple-digit increases in loans that were originated as a result of the experiments and were able to effectively scale and embed those new messaging techniques as part of the business-as-usual processes.