How a mid-size private bank earned USD 2.4 Mn by targeting offers by segment, not by product

01At a glance

02The challenge

We marketed to our 2.1 million retail customers in three broad groups: salaried, self-employed and senior citizens. Every campaign went to everyone in a group, so response rates were low and customers tired of offers that did not fit them. Branch teams had their own ideas about who to call, but no data to back them. Meanwhile, customers with room for more products were taking them elsewhere, and dormant accounts kept rising. We needed to know who our customers really were and what each group was likely to buy next.

03What we did

We brought three years of account, transaction, product and campaign data together on ALTeX Hub, inside the bank’s own environment, with personal data masked. Our data scientists built behavioural segments from how customers actually earn, save and spend, then trained propensity models for the six products that mattered most. We tested the segments with branch and product teams before launch and set up control groups, so every campaign could be measured. Two data scientists, a data engineer and a banking analyst worked with the marketing team for ten weeks. The first targeted campaigns went live in week six, and the models now refresh every month.

04What we found

05The result

• USD 2.4 Mn earned
• First targeted campaigns live in 6 weeks
• 100% of 2.1 million retail customers segmented, against 3 broad groups before

06In their words

07What’s next

The bank now refreshes its segments monthly and has asked us to model pricing for its credit-card book.