Our client, a domestic commercial bank, was launching a new retail product and needed to identify the right customers to target from among their hundreds of thousands of existing customers. The goal was a focused campaign — not a broadcast.
The problem
The bank had already filtered its retail customer base using internal data: account balances, banking behaviour, and home addresses. The resulting list was still far too large for an efficient postal and telephone campaign to be viable. The number of customers to approach had to be reduced further, but without simply cutting the list arbitrarily — the reduction needed to be based on who was genuinely likely to respond.
What we did
We analysed the profile of each address in the filtered list at door level, using our demographic databases. For every address, the analysis covered:
- The demographic composition of the household
- Living situation and property characteristics
- Estimated income level
- Overall living conditions and consumption behaviour
This profiling allowed us to rank every address by its propensity to respond to the specific product offer — and to cut the list to those households where the characteristics matched the product most closely.
The outcome
We successfully reduced the number of targeted customers to one-sixth of the original list.
The impact was significant across three measures:
- Campaign cost fell substantially, as postage, printing, and call-centre time were concentrated on far fewer contacts
- Success rate rose from the typical 2–3% to above 12%
- Profitability on the new product increased considerably, because the cost of our service was a small fraction of the potential waste coverage cost of a “full” untargeted campaign
The engagement demonstrates the commercial value of address-level data when applied to an already-qualified customer base — the insight was not in identifying who might want the product, but in identifying where those people actually live.