In Customer Management Level I, we discussed the different segmentation techniques that banks have used to segment their customers into groups that they can then target with offers.

The process is:


Typically, banks have hundreds of thousands or millions of customers, all using their products and services in slightly different ways, generating millions of datapoints that, if understood would help the bank allocate its finite resources to create value.

In this module we will now focus on how banks use behavioural segmentation to understand where to allocate resources to acquire new customers, grow existing customers and keep customers from moving to a competitor – asking questions to develop what they need to do to direct resources to where value will be created.

Segmentation Process

This begins with the selection of a suitable base. Depending on the purpose of the segmentation exercise, it could be the entire population of adult consumers, or a sub-set of them, or their current portfolio of customers. Marketers are looking for a means of achieving internal homogeneity (similarities within segments) and external heterogeneity (differences between segments) to minimise the differences between members of a segment and maximise differences between each segment.

The process must yield segments that are meaningful for the specific purpose. It requires a good understanding of the market to be segmented. Segments must be:

  • Identifiable – the extent to which managers can identify or recognise distinct groups within the portfolio or marketplace;
  • Substantial – determines whether the segment or group of customers is sufficient a size to be profitable, in terms of numbers or purchasing power;
  • Accessible – analyses whether marketers can reach the targeted customers with promotion efforts;
  • Responsive – determines if the consumers or customers within the defined segment respond to marketing offers targeted at them;
  • Actionable – segments are ‘actionable’ when they provide guidance for marketing decisions.

In undertaking the above, the bank may determine that some segments in its portfolio of customers aren’t currently actionable or valuable. It’s not uncommon for senior executives and others to demand that the bank withdraws from or closes certain customer segments. However, the fact that the segment may not be valuable now isn’t an automatic reason to close or withdraw, but a challenge for marketers to find a way that turns these customers into valuable ones.

Questions Executives Should Ask

Instead, executives should ask how the bank can reverse the situation by allocating resources to where customers are asking for solutions to problems, help in achieving their dreams or where the segment is currently losing money.

We discuss the subject of how to allocate resources in more detail later in this module.

Behavioural Segmentation

In Customer Management Level I, we described behavioural segmentation as “Using behaviour exhibited towards your organisation, include the products and content consumed, and interaction frequency.”

Customers with the same set of products, using the same channels can behave differently.

Ideal Customer

The bank usually has an idea who their ideal customer should be. It sometimes comes as a shock that they don’t have the customers they thought they did. This may explain why they aren’t meeting their strategy or plan and will lead to a radical rethink to deal with the reality.

Data Availability and Quality

The earlier section on Customer Data describes the typical data that a bank must collect and make available for segmentation and other analysis. The data used must cover a reasonable period, for example a six months to twelve months period.

Data quality is crucial. However, it is usually at this point that the bank realises it has data quality issues that must be fixed before behavioural segmentation can be undertaken.

Behavioural and Other Data Used

Depending on the purpose of the segmentation, this will include:

  • Demographic data: such as age, marital status, dependents, occupation;
  • Length of time as a customer;
  • Products held by type: current/checking, savings, loans, mortgages, insurance;
  • Product usage: light user; heavy user or moderate user;
  • Transactions by type: total credit and debit transactions categorised by card transactions, recurring payments (direct debits, standing orders), electronic payments – inwards and outwards, over-the-counter transactions;
  • Channels used by type: branches, ATMs, telephone banking, Internet banking, mobile banking;
  • Channel usage/preference: by channel – light user; heavy user or moderate user, how it is used – browsing, transacting, interacting;
  • Channel adopter status: early adopter, late adopter, laggard;
  • Customer current value;
  • Customer future value (if available);
  • Life cycle and event information: gathered from customer;
  • Contact preferences: how they prefer to be contacted by communications channel and response preference – how they respond;
  • Buyer status: product or service in progress;
  • Dormancy (no activity within last 30 days): dormant, not dormant;
  • Delinquency: delinquent borrowing, no delinquent borrowing;
  • Current proposition: if appropriate;
  • Current segment: if appropriate.

Customer Data Analysts and Customer Analytics Platform

We have discussed the capabilities of both the analysts and their platform. Both must be capable of handling more advanced data mining, integration and analysis techniques using machine learning, artificial intelligence, and advanced visualisation techniques.