Woman holding credit card while using laptop

Case study October 1, 2024

From monthly batches to real-time moments: rebuilding a bank's marketing engine

Real-time engagement, built to match how customers move.

Service

Adobe

Industry

Financial Services

Context

Ten minutes. That's roughly how long a customer in Latin America gives a digital channel before walking away, and more than six in ten will do exactly that if the response doesn't come in time. E-commerce and delivery apps set that bar. Banks are now held to it too.

For a bank serving both retail and business customers, that bar applies to everything: every marketing message, every one-time password, every fraud alert must move at the speed of the channel it travels on.

One of Argentina's largest private-sector banks felt that pressure directly. It serves millions of retail and business customers nationwide, and over several years had built a broad marketing technology stack that included Adobe Campaign Standard, Adobe Audience Manager, Adobe Target, Adobe Analytics, and Adobe Experience Manager. That stack carried the bank through its first wave of digital marketing maturity. But its batch-oriented core was reaching a ceiling right as instant, event-driven communication stopped being a nice-to-have and became the baseline customers expected.

The problem

As the bank grew, so did what it asked of its marketing engine. Adobe Campaign Standard had carried the bank's marketing operations through years of expansion, but the volume and immediacy the business now needed had outgrown what a batch-oriented tool was ever asked to handle. One-time passwords for security-critical actions ran through the same batch pipeline as marketing campaigns, competing for a shared sending quota and creating real operational risk whenever demand spiked. Push notifications followed a similar batch cadence, so the bank couldn't react to what a customer was doing in the app in the moment that mattered. Frequency and prioritization rules were coarse and largely batch-executed too, which made it hard to protect customers from over-contact, especially once the bank needed different capping logic by day, month, product, and channel, all at the same time.

Underneath all of this sat another challenge: the bank couldn't see its own customers clearly. Retail and business profiles lived in separate structures. Delivery and tracking data was aggregated rather than granular. And there was no reliable way to hold out a control group and measure what a campaign changed, versus what would have happened anyway. None of this was a shortcoming of the platform itself. It was the natural result of the bank's own growth: as the business evolved, the technology behind it had to evolve too.

The solution

A multi-year partnership with Nortal moved the bank's marketing and customer data foundation onto Adobe's real-time stack: Adobe Experience Platform, Real-Time CDP, Adobe Journey Optimizer (AJO), and Adobe Customer Journey Analytics (CJA), gradually retiring Adobe Campaign Standard and Adobe Audience Manager as the new architecture took over. We built it in stages, starting with the channels that mattered most to the customer experience.

Every channel had to move at the speed of the customer, without the bank losing ground it had already earned. Email moved onto AJO on dedicated subdomains, so the bank kept full control over its own domain governance, then earned a stable sending reputation through a 30-day warming program before scaling volume up. One-time passwords, previously stuck in the same batch queue as marketing email, moved to real-time transactional sends with no dependency on sending quotas, removing a source of operational risk during high-demand periods. Push notifications moved onto Adobe's mobile SDK and native push services, so the bank could react to what a customer was doing in the app right then instead of waiting for the next batch. For SMS, we built a custom integration with the bank's existing, already-approved provider, keeping its established compliance and monitoring relationship intact while adding real-time, API-based delivery for both transactional and marketing messages.

With the channels rebuilt, the bank needed to protect customers from over-contact and prove what its campaigns were achieving. Native business rules in AJO now manage daily and monthly frequency caps at the profile level, while a complementary model using profile attributes and audience segmentation handles weekly caps by product and channel, something no single native rule could do on its own. We designed a control group feature: a combination of profile attributes, journey splits, and audience logic that reliably holds out a comparison group and tracks it through to CJA, so the bank can finally measure the lift a campaign creates instead of assuming it.

The bank also needed to see its own customers clearly again, across every channel and business line. Delivery and tracking logs were extended to unify individual customer, device, and business-account identifiers into one dataset, giving the bank one-to-one traceability across its full customer base. A shared content fragment now centralizes the preference center link across every email template, so one update replaces what used to be manual edits across many templates. A business calendar dataset, combined with dynamic lookups inside journeys, keeps regulated communications within business hours without the tangle of extra wait steps that used to be required. And in the bank's insurance journeys specifically, replacing manual validation steps with exit criteria and reaction events cut the number of nodes per journey by about a quarter, making them easier to maintain and leaving room to add new use cases.

Finally, the bank needed all of this to add up to one connected view of the customer, not another disconnected data source. CJA pulls together data from the bank's CRM, its web and mobile SDKs, and AJO itself, giving teams across the bank one view of the customer journey, individual and business alike, across online and offline touchpoints, without waiting on a manual data pull.

The impact

The shift from batch to real time shows up first in reach and scale. 100% of the bank's release volume now runs through the new, real-time infrastructure, and the bank can reach its entire credit-card customer base in real time.

Insurance journeys are roughly a quarter simpler to maintain after the node-reduction work, which frees capacity for new campaigns instead of more workarounds. One-time passwords and push notifications now move at the speed of the transaction itself, cutting the delays that used to create friction for customers and reducing the operational risk of high-demand periods.

Most importantly, the bank now has something it didn't have before: one unified view of every customer, retail or business, across every channel, and a reliable way to prove what a campaign changed instead of assuming it. Accurate identity, real-time delivery, and honest measurement, together, are what turn a marketing stack from a cost of doing business into a growth engine.

100%

of release volume now running through real-time infrastructure

100%

of credit-card customers reachable in real time

25%

fewer nodes in the bank's insurance journeys, freeing up room for new use cases

unified view of every customer and channel, individual and business alike

Batch logic doesn't announce itself as a problem. It just quietly costs you the moment: the OTP that arrives too late, the push notification that fires after the customer already left the app, the campaign you can't prove worked.

If that sounds familiar, we'd be glad to walk you through what a real-time architecture could look like for your platform, starting with the piece that's costing you the most right now.

Get in touch