Retail Banking: Insights into customer loyalty

Retail Banks work in complex, changing environments so research needs to be agile to keep pace with new opportunities and risks. One of the great benefits of social insight is its flexibility and responsiveness to sudden changes in research priorities. In this case, a bank poised to reduce interest rates needed to understand the potential risk to its customer base.

What happens when a bank concludes that the interest rate of its flagship savings accounts is not sustainable and a rate cut for customers is unavoidable? When rates fall, customers are more likely to look for alternatives, so the bank has to make a careful calculation between the improvement in its margin and the potential cost of losing customers.

The cost of new customer acquisition heavily outweighs the cost of retention. Even the most conservative estimates suggest it costs between 7 and 12 times as much to recruit a new customer than to retain an existing one. It obviously makes good business sense for banks to keep their customers contented and on-board. 

We cannot answer the question with the sentiment scores common in social media reporting. A rate cut for savers is a BAD THING. I think the bank can work that out for itself. In this instance, understanding the underlying attitudes of the customer base and the incremental risk to loyalty is central to quantifying the impact of the operational decision. 

This means examining how, when and where customers engage with the rate change, both in terms of what they say and what they do.


Our first task was to analyse historic customer activity in the 12 months prior to the rate change. There is a surprising amount of social discussion about personal banking, much of it centred on day-to-day concerns such as how to open an account, where to go with a question etc. This content may not grab headlines, but it was all relevant to our purpose of measuring the frequency and strength of customer engagement, based on unsolicited attitudes and observed behaviours.

Insights into customer intentions

The initial analysis created benchmarks which we used to measure deviation from established patterns of activity once the rate change was announced. By measuring these changes in behaviour, we divided the customer base by four levels of risk. The composition and momentum of each group was tracked before and after the rate change, to analyse the impact of the bank’s action.



The segmentation was designed to measure the reaction and subsequent actions of each group, based on a hypothesis that highly-motivated customers would be at greatest risk of leaving the bank.

I’ll say it again, our objective was not to measure sentiment. There was a lot of noise generated by the move, much of it hostile. But the purpose of our study was to see through the immediate reaction and to detect a palpable intention to leave. We all say things in the heat of the moment. 

Over the duration of the study period there were significant changes in the composition of each risk segment. We observed customers reacting to the news, followed by a period of diminishing activism by a majority of customers. 


So what did this mean for risk? A stable signal of customer intention appeared over time, as we constantly reassessed opinions and behaviour. The highest level of risk was assigned to customers who exhibited uncharacteristically high levels of activity (compared to their previous online behaviour) and who consistently stated an intention to leave. Based on the evidence of the analysis, a hard intention to take action stabilised at 8 per cent of the sample after 10 days.

This doesn’t mean that 8% of clients will leave. There are other variables to consider, not least the availability of alternative offers from rival banks. Some of those customers looking to leave might be disappointed by what’s on offer elsewhere and decide to stay (which gives me an idea for a useful follow-on analysis).  But the outcome gave us a fixed point to assess current customer intentions against the historic rate of customer churn. The end result is a better understanding of the increase in risk.


Blue trend line (left axis): % of customer sample expressing intention to leave.
Red trend line (right axis): strength of intention to leave. Scale 1 (benign) to 5 (leave).
The average level of intention strengthens within the Leavers group as less motivated customers drop out of the high-risk category. So we see the emergence of a minority group of motivated Leavers.
After 10 days Motivated Leavers stabilise at 8% of the sample.


Postscript,16 Jan 2017

We have taken another look at this case study to see what happened over the longer term. A general assessment of the level of account holder activity shows that most customers have resumed normal levels of engagement with the bank and with each other. There is less motivation to move, but there is also less certainty about the quality of the returns available from the bank, which leaves customers more open to rival offers and more sensitive to other aspects of the relationship with their bank, such as quality of service, responsiveness, convenience and flexibility.




Have your say