Best use of FastStats® 2016 Award Finalist
Using FastStats Modelling™ to understand customer behaviour types and what product should be offered next
Since 1853, Skipton Building Society (SBS) has been dedicated to helping people save for the future and achieve their dreams of owning their own homes. They have come a long way, and are now one of the UK’s largest building societies. Although they have grown and changed, they have always stayed true to their roots, with the core value of ‘being there for our customers to help them have a better life is still – and always will be – at the heart of what we do’.
Based in North Yorkshire, they now have 95 branches across the UK and customers are able to interact with SBS in branch, online and telephone. SBS always aim to put the customer first. With no shareholders to answer to, their time, energy and expertise goes into looking after their customers. This is demonstrated by the approach of listening to what their customers say and acting upon it, this customer feedback helps to develop accounts and services that have real value to their customers’ life plans.
In 2015 SBS were recognised by KPMG Nunwood as the 7th best company in the UK for ‘Customer Experience Excellence’, as well as being one of ‘The Sunday Times 100 Best Companies to work for’ in both 2015 & 2016.
SBS are relentlessly focused on providing excellent customer experiences. This objective runs throughout their business and has driven numerous strategic programmes at the building society. SBS chose Celerity as their strategic partner to support them with marketing objectives around a framework of the right product, to the right customer, at the right time, with the right message.
FastStats supports the team with various marketing initiatives and is used to generate the actionable insight required for the team’s customer-centric strategy.
On top of the SCV, Celerity implemented leading analytics platform, Apteco FastStats, and Adobe Campaign for campaign management and execution. This new marketing stack provides SBS with the capability to make informed decisions about their customers, action the insight gained and automate multi-channel, highly personalised cross channel marketing communications.
One of the key business challenges the new solution needed to help SBS overcome was the retention of savings customers. The building society were seeing drop off especially prevalent for customers who had defined term savings products such as a 5 Year Fixed Rate Bond - at the point of account maturity, customers can easily take their funds to another bank or building society.
SBS worked with Celerity on a ‘Next Best Savings Product’ project in order to attempt to overcome this key challenge by understanding what product might be a suitable next recommended product to customers who looked like they were about to churn. The project included the creation of a number of predictive models and then 21 PWE predictive scores for every customer on their individual likelihood to take up different types of savings products. The model scores update daily for use within FastStats and Adobe Campaign.
The Next Best Savings Product Project: Retention of Savings Customers
Celerity and SBS worked together to understand the challenges around savings accounts and prioritise initiatives in line with expected positive impact and the investment vs return. The top business challenge agreed was the retention of savings customers; especially those with a term account e.g. 3 Year ISA, and those customers considered younger savers (under 40s). This challenge was broken down into three key parts to ensure the solution created would be robust, long-lasting and impactful. This highlighted three areas where further insight was required to help reduce churn and increase customer LTV:
- Early Attrition: it was highlighted that a significant portion of new account starters had a tendency to close their account notably early after starting the account. The reasons identified as to why this happened included ‘people who chase after the best rates’ and ‘people who may open a savings account as a pre-requisite for a mortgage account’. Insight was required to identify the individuals likely to attrite early so that SBS could engage them before they leave.
- Customer Behaviour Type: It was discovered that customer behaviour type is a key factor in the retention of savings customers. Their attitude and decision-making process impacts their likelihood to behave a certain way. SBS needed insight into the behaviour type of their customer base so that they could communicate with them in a more effective manner.
- Next Best Savings Product: SBS needed insight into the likelihood of customers to take out a next product as well as when and how to engage them most effectively. This is the key challenge because of the numerous options available to customers in the financial services market.
Following this understanding of the insight required, Celerity and SBS planned to create multi-faceted customer models using Apteco's FastStats software with the aim of providing actionable insight that could transform the communication journeys that savings customers go on. The ultimate aim was to align around the 4 Rs: Right Product, Right Time, Right Customer, and Right Message.
The variables used in the process included, transactional, behavioural, demographic, attitudinal and geo-demographic. Celerity also utilised qualitative data in the savings behaviour research to incorporate industry research into the data strategy. The volumes used in the analysis selection were statistically significant with 95% margin of confidence and 0.3% margin of error against the Base selection.
The project was broken down into four key steps:
- Fight Early Attrition
- Understand Savings Customer Behaviours
- Next Best Savings Product Modelling
- Test and Roll-Out
Step 1: Fight Early Attrition
Strategy: Use insight to predict the customers who may close their account early, and engage them so they stay with SBS for longer.
The FastStats Profile tool was used to analyse the most indicative characteristics of this behaviour. 28 variables were used across the nine savings product groups. The ‘Mean Index’, ‘Information Gain’ statistics and feedback from the SBS team were then used to make decisions about the best variables to action against.
Step 2: Understand Savings Customer Behaviours
Strategy: Use insight and research of savings behaviours to predict SBS customer attitudes and market to them more effectively.
The research highlighted three key behaviour strands:
- Some savers have a preference to ‘remain with the same’ type of savings product that they currently have.
- Some savers have a preference to have ‘variety’ in types of products, spreading risk.
- Some savers have a preference to ‘move to a better’ type of product (better interest rate return).
These identified behaviours were presented to a research panel with varying savings attitudes to explore the reason for those views, thoughts and feelings. The output of the panel workshop reinforced the need to apply attitude into a savings customer retention programme. The workshop highlighted that the decision-making process of the varying attitudes are extremely different, therefore a single retention programme strategy for all customers would not be appropriate.
The FastStats Profile tool was used to create three scores for each customers likelihood to fit into each of the identified savings behaviours. The variables that were used help to tell a customer story leading up to the point whereby they are likely to want a new savings account. This insight and understanding can be used to drive the SBS communication strategy to promoting new savings products to a customer at the right time. The image below looks at the variables that can be used to determine the best communication and journey at certain relationship points.
The finalised PWE models had significant uplift based on Model Power against random, and also against a holdout test segment. The final behaviour model variables were then used as part of the Next Best Savings Product modelling process in Step 3.
Step 3: Next Best Savings Product Modelling
Strategy: Understand and promote the right choice of savings product to a customer at the right time.
The final profile reports showed that all three savings behaviour models proved to be very strong predictors of identified customers. For example, customers who are significantly likely to want a “Long Term” savings account have the following characteristics:
- Remain saving behaviour attitude
- Change to a ‘Long Term’ account from another long term account
- Change accounts with one to two year frequency
The final predictive model scored and grouped customers into 20 semi-deciles. The ‘Profile Model’ wizard was used to create 20 quantile categories selector variable, ensuring there are equal volume of records per category. The developed propensity models had model power ranging from 0.50 to 0.71, indicating that the models can significantly identify customers deemed as very likely to want the specified type of savings product, as well as predict customers who are significantly likely to look like them.
Step 4: Test and Roll Out
Strategy: Use the ISA Next Best models as a test against traditional selection during ISA Campaign Season.
The ISA versions of the ‘Next Best Savings’ models (ISA Long Term, ISA Short Term and ISA Instant Access) were tested for 2016 ISA Season campaigns.
The test campaign went out to 192,000 customers across email and direct mail. SBS achieved an overall product uptake response rate of 3.9% (3.8% for models based work flow). Overall 98% of all respondents were identified in the best model deciles and 80% of the contacts were identified in the best model deciles.
The predictive models are now being used to drive customer journey planning to ensure the right product, to the right customer, at the right time, with the right message.