Case Study: BNP Paribas achieves real-time targeted offers and increased revenue with Hazelcast

How BNP Paribas Bank Polska uses an event-driven architecture to increase revenue opportunities

BNP Paribas faced a marketing challenge: present timely, relevant product offers to customers right when they needed them, rather than through slow, batch-driven campaigns. The bank’s existing SOA/ESB and CRM/data-warehouse processes could take up to two days to generate offers, so BNP Paribas turned to Hazelcast to add real-time stream processing and in-memory enrichment to its infrastructure.

Hazelcast was deployed as a streaming engine that plugged into the bank’s publish/subscribe bus, enriching events from an in-memory account/product repository, running Python-based promotion rules, aggregating transactions, and publishing consolidated records back to the bus. The solution delivered sub-120 ms processing, supported about 10 million events/day across two multi-node clusters, and enabled ~70 campaigns; offer conversions increased four-fold, the initiative became profitable within six months, and the pilot cost about one-tenth of a data-warehouse approach.

BNP Paribas

Szymon Domagala

Enterprise Architect