Reinventing Bank Loyalty through Decision Intelligence — 7 Takeaways

From the Consumer Bankers Association (CBA) webinar “How Decision Intelligence Can Reinvent Bank Loyalty,” presented on August 19, 2026, by Curinos’ Sarah Welch and Olivia Hamel.

1. The single-bank customer is disappearing, and fintechs are moving upmarket.

Low switching costs, fee-free entry points, and instant transfers have eroded the inertia that used to hold banking relationships together, and the data shows it: the average number of checking relationships per customer rose from 1.5 in 2019 to 2.2 in 2025, a 47% increase in just six years. At the same time, fintechs that historically targeted paycheck-to-paycheck customers are moving upmarket with genuinely competitive offers. Chime+ offers 3.75% APY for direct deposit with no fees or balance minimums, and SoFi Plus offers 4.50% APY and other member-only benefits for a $10/month subscription. That combination squeezes banks on two sides at once: existing customers have less reason to consolidate, and new entrants are courting the exact segments banks have historically treated as core. Rather than a marginal shift, this is a structural change in the competitive landscape that traditional loyalty approaches were not built to handle. Banks that don't adapt their loyalty strategies risk losing share of wallet among customers who technically still bank with them.

2. Loyalty has a measurable, quantified payoff across the customer lifecycle.

Loyalty manifests in three observed outcomes tied to specific points in the customer lifecycle. First, banks perceived as rewarding customers see a 1.75x higher acquisition/purchase rate compared to other banks, giving loyalty a real role at the top of the funnel, not just with respect to retention. Second, once a relationship is established, primary customers in loyalty programs bring 3-4x higher balances than non-primary customers. Third, achieving primacy is linked to 60% lower front-book attrition (customers on book less than 90 days), meaning the loyalty investment pays off earliest exactly when relationships are most formative. Clearly, loyalty isn't a cost center, it's a lever with quantifiable returns across acquisition, deepening and retention simultaneously.

3. Traditional loyalty tiers see only a snapshot, which masks how customers are evolving.

Good-better-best tier models work by evaluating customers at a single point in time. But because 80% of customers make up only 14% of total deposits under a typical tiered construct, the vast majority of the customer base is invisible to premium recognition. In a tiered system, three customers holding $80,000 today would look identical. In fact, one might have held that balance for over a decade with no growth, one might have moved it from another bank just last week, while the third may have saved it gradually from just $2,000 through years of discipline. A snapshot-based system treats all three identically, even though their trajectories — and their future value to the bank — are completely different. Rewards need be defined by relevant recognition, guidance, and activation reflecting a customer's individual trajectory, not merely their current balance. This reframes the entire premise of tiered loyalty: point-in-time balance is a weak proxy for where a relationship is headed.

4. The often-overlooked mass market segment carries real, quantifiable growth potential.

Over a five-year period, mass market balances (customers under $10K at three months on book) grew from just 10% of a portfolio's balance mix in year one to 34% in year five. Their balances grew by 5.4x, while at the same time mass affluent customers' average balances declined to 0.7x of their original deposit. That means the segment that a tiered system ignores today is disproportionately the segment that grows fastest if a bank retains and nurtures it. This reframes entry-level and mass-market customers from a cost center to a growth engine hiding in plain sight. 

5. Tiers, triggers, and decision intelligence answer three genuinely different questions.

Three concepts may be used interchangeably, but they serve very different purposes. Tier logic answers “do they qualify?” It's a point-in-time, segment-level determination of benefit eligibility, and its operating model is static. Trigger stacks answer “did something happen?” They’re event-level and reactive, signaling readiness to engage without judging whether engaging is actually the right call. Decision intelligence asks a different question altogether: “what should happen next?” It evaluates at the individual customer level, dynamically, against the bank's actual objective (primacy, retention, funded balances, cost of funds). It isn’t intended to replace tiers or triggers — the best loyalty programs use all three together — but it acts as the coordinating layer that determines whether a customer’s change or a trigger event deserves an action.

6. Concrete customer examples show that “no action” can be the right decision.

Avery, a 28-year-old checking customer with just $100 in savings would generate zero recognition under a tiered system because she's below every threshold. A trajectory-based view, on the other hand, catches her setup for recurring savings, her emergency fund milestone, and her income increase, delivering a relevant nudge or proactive guidance at each step rather than leaving her undetected and ignored.

7. Blasting more outreach can backfire — and Curinos One is built to fix the coordination gap, not the data or execution.

Because historic loyalty and marketing efforts have defaulted to acting on every eligible trigger, customers have been inundated with outreach they experience as irrelevant rather than rewarding. Almost half cite irrelevant offers as the most bothersome form of bank outreach, followed closely by generic, non-personalized content (43%) and being overly sold to (42%). This isn't just an annoyance; it actively erodes trust and can make customers less receptive to future, more relevant interactions, thereby raising the cost of every subsequent touchpoint.