Next Best Action vs. Product-Based Cross-Selling: The Paradigm Shift in Banking

From Product Catalog to Individual Recommendation

Product-based cross-selling follows a simple logic: every product in the portfolio gets a target segment, which is then systematically approached with matching offers — customers without a credit card get a credit card offer, customers without a brokerage account get a securities offer. Next-best-action (NBA) flips this logic on its head: the starting point isn't the product, but the individual customer, for whom a model determines — across the entire product range — the action with the highest expected value. That action might even be no contact at all, a service touchpoint instead of a sales pitch, or a human advisory conversation.

The Structural Weaknesses of Product-Based Cross-Selling

Product-based cross-selling almost inevitably creates collisions: when five product teams independently plan campaigns for their own target segments, a customer who qualifies for several products may receive multiple offers in quick succession — or contradictory messages. The product-based approach also fails to account for which offer is genuinely most relevant and likely to succeed for that specific customer. A customer might formally qualify for both a credit card upgrade and a home savings product, yet only one of the two carries a realistic conversion probability above 5%.

How Next-Best-Action Models Decide

NBA systems calculate, for each customer and each possible action in the product catalog, a success probability along with an expected value contribution (conversion probability multiplied by product margin or customer lifetime value). Contextual factors are layered in as well: was the customer already contacted in the last 30 days? Is there an active complaint that rules out a sales approach? The system then selects, per customer, the action with the highest overall value — not necessarily a sales offer, but potentially a service reminder, or no contact at all if the expected benefit is too low or the risk of a negative reaction too high.

Quantifiable Effects

Banks switching from product-based cross-selling to next-best-action typically report significant improvements: overall conversion rate across all campaigns often rises 20–35%, while contact frequency per customer drops 15–25% at the same time, because fewer but better-matched offers are sent. Complaint rates tied to marketing outreach — a frequently underestimated cost factor — often fall 30–50%, as irrelevant or redundant offers disappear.

Prerequisites for Making the Switch

The shift to next-best-action requires an organizational change that's often underestimated: product teams accustomed to owning their own campaign budgets and targets must fit into a central prioritization system that doesn't treat every product equally but decides based on customer value. This can create internal conflicts of interest when a product team has an aggressive sales target but the NBA model recommends a different product for most addressable customers. A clear governance structure with overarching KPIs — total customer value rather than single-product sales — is necessary to resolve these conflicts.

Data Foundation and Model Quality

The quality of an NBA system depends directly on its underlying data foundation — particularly transaction data that reflects actual customer behavior rather than just demographic or product-holding attributes. Equally important is a working signal hierarchy and ongoing model monitoring, to ensure recommendations stay relevant over time and aren't skewed by outdated patterns.

Regulatory Considerations

From a GDPR standpoint, it matters that automated product recommendations remain explainable — especially when they determine customer communication without human involvement. Documented decision logic explaining why one product was recommended over another matters both for internal audits and for regulatory inquiries under BaFin supervision.

Change Management as a Critical Success Factor

Shifting to next-best-action doesn't just change systems — it changes roles and incentive structures. Sales leaders whose bonuses were previously tied to selling a specific product need to adjust to overarching customer-value metrics that bear less directly on their own product line. Without accompanying change management, this often creates resistance that undermines an otherwise technically sound NBA system in practice, as advisors or campaign managers quietly ignore or work around the recommendations.

Successful institutions support the rollout with transparent communication about how performance targets are changing, and with pilot phases where sales teams see, through concrete examples, that NBA recommendations improve their own success rate compared with previous practice. A German regional bank taking this path should plan for a transition period of roughly six to nine months before sales teams implement NBA recommendations in customer conversations without significant deviation.

Conclusion

Next-best-action replaces a product-centric logic with a customer-centric one. The switch delivers higher conversion rates and less wasted reach, but demands an organizational realignment — from product teams optimizing in silos to a central, customer-value-driven prioritization.