Revenue & Product Recommendation Engine¶

Next-best-product recommendations, fee optimization, and product bundling analytics for retail and commercial banking.

Priority: P3 — Operational Excellence
Time to Value: 10-12 weeks
Category: Revenue & Product Strategy

Business Problem¶

Banks have extensive product portfolios — deposits, loans, cards, insurance, investments, trade finance — but most customers hold only 1-2 products. Revenue growth through existing customers is significantly cheaper than new acquisition, yet banks struggle to cross-sell effectively:

Low product penetration — average customer holds 1.8 products despite eligibility for 4-5, leaving substantial fee and interest income unrealized
Generic campaigns — mass marketing campaigns with 2-3% conversion rates because offers are not personalized to individual customer context and needs
Fee income leakage — pricing structures not optimized by segment; high-value customers receive unnecessary waivers while price-sensitive segments are overcharged
No competitive intelligence — pricing decisions made without visibility into competitor rate offerings and market positioning
Bundle blindness — products are sold individually rather than as value bundles (e.g., salary account + credit card + personal loan + insurance) that increase stickiness and lifetime value

Capabilities¶

Next-Best-Product Recommendation¶

AI model predicting the highest-propensity product for each customer based on their current holdings, transaction behavior, life-stage signals, peer cohort patterns, and channel preferences.

Dynamic Pricing & Fee Optimization¶

Segment-level pricing optimization for interest rates, processing fees, annual charges, and service fees — balancing revenue maximization against customer retention and competitive positioning.

Product Bundle Design¶

Data-driven identification of product combinations that maximize customer lifetime value and reduce churn. Design and test bundles for specific segments (salaried professionals, SME owners, retirees).

Campaign Effectiveness Analytics¶

Measure and optimize marketing campaigns: A/B test different offers, channels, and messaging; attribute revenue to campaigns; and continuously improve targeting models based on outcomes.

Competitive Rate Intelligence¶

Monitor competitor product rates and offers from market data and public sources. Alert product teams when competitive gaps emerge and recommend pricing adjustments.

Data Sources & Ontology Mapping¶

flowchart LR
    subgraph Data Plane
        CBS["Core Banking System"]
        SFSC["Salesforce FSC"]
        MKT["Market Data & News"]
    end

    subgraph Ontology Entities
        CUST["Customer Profile"]
        HOLDINGS["Product Holdings"]
        TXN["Transaction Patterns"]
        INTERACT["Interactions & Campaigns"]
        RATES["Market Rates & Offers"]
    end

    subgraph AI Workflow
        NBP["Next-Best-Product Model"]
        PRICING["Pricing Optimizer"]
        BUNDLE["Bundle Analyzer"]
        CAMPAIGN["Campaign Analytics"]
    end

    CBS --> CUST
    CBS --> HOLDINGS
    CBS --> TXN
    SFSC --> INTERACT
    SFSC --> CUST
    MKT --> RATES

    CUST --> NBP
    HOLDINGS --> NBP
    TXN --> NBP
    INTERACT --> CAMPAIGN

    NBP --> BUNDLE
    RATES --> PRICING
    HOLDINGS --> PRICING
    CAMPAIGN --> NBP

Ontology Entity	Source System	Key Fields
Customer Profile	CBS + Salesforce FSC	CIF, Segment, Age, Income Band, Occupation, Tenure, KYC Status
Product Holdings	Core Banking System	Account Type, Product Name, Balance, Rate, Open Date, Status
Transaction Patterns	Core Banking System	Category (salary, rent, shopping, travel), Frequency, Average Amount, Channel
Interactions & Campaigns	Salesforce FSC	Campaign ID, Channel, Offer, Response, Conversion, Date, Advisor
Market Rates & Offers	Market Data + Competitive Intel	Competitor, Product, Rate, Fee, Promotion, Effective Date

AI Workflow¶

Customer Feature Assembly — Build a comprehensive feature set per customer: demographics, product holding vector, transaction category distribution, channel preference, tenure, and relationship depth
Propensity Modeling — Train multi-label classification models predicting purchase probability for each unowned product category; rank by expected revenue contribution
Life-Stage Detection — Identify life-stage triggers from transaction patterns: salary increase (income jump), marriage (joint account activity), home purchase (large debit + property search), retirement (pension credits)
Pricing Simulation — Model price-response curves per segment: how does conversion change at different fee/rate levels? Identify revenue-optimal pricing for each segment × product combination
Bundle Optimization — Association rule mining + CLV modeling to identify product combinations that maximize 3-year customer value; design bundles with pricing incentives for adoption
Campaign Orchestration — Feed recommendations into Salesforce FSC for advisor-led sales and digital campaign platforms for self-service channels; track conversion and attribution
Output — Product recommendation dashboard for branch and digital channels; pricing recommendations for product teams; bundle proposals for marketing; campaign performance for CMO

Dashboard & Alerts¶

Key Metrics¶

KPI	Description	Target
Product Penetration	Average active product categories per customer	> 3.5 (up from 1.8)
NBP Conversion Rate	% of next-best-product offers that convert to sale	> 15%
Fee Income per Customer	Annual fee income per active customer	12% year-over-year growth
Campaign ROI	Revenue attributed / Campaign cost	> 8x
Bundle Adoption Rate	% of new customers taking bundled vs. single product	> 40%
Price Competitiveness Index	Bank's rate positioning vs. top 5 competitors per product	Within 25 bps of market median

Alert Rules¶

Alert	Trigger	Severity	Action
Competitive rate gap	Competitor launches deposit rate >50 bps above bank's rate for same tenor	High	Notify product team; model attrition risk and pricing response options
Campaign underperformance	Campaign conversion rate <50% of predicted target after 2 weeks	Medium	Pause campaign; analyze targeting model; adjust offer or audience
High-value opportunity	Customer with CLV >$50K identified with 3+ unowned high-propensity products	Medium	Push priority recommendation to assigned advisor in Salesforce
Fee waiver anomaly	Branch-level fee waivers exceed 2x average for segment	Medium	Flag for audit; review waiver authority and justification
Product attrition signal	Product category closure rate exceeds 5% monthly for any segment	Info	Investigate root cause; assess competitive positioning

ROI Model¶

Metric	Before	After	Impact
Product penetration	1.8 products/customer	2.8 products/customer	56% increase in product depth
Cross-sell conversion	3% campaign conversion	15% AI-targeted conversion	5x improvement
Annual fee income	$42M	$47M	$5M incremental fee revenue
Customer attrition (multi-product)	12% for single-product holders	4% for 3+ product holders	Stickiness multiplier from bundling
Campaign waste	40% of budget on low-propensity targets	12% waste rate	70% reduction in ineffective spend

Estimated Annual ROI

$5M - $10M annually from incremental fee income, improved cross-sell conversion, reduced attrition, and campaign efficiency — across a mid-size retail bank with 500K+ customers.

Implementation Notes¶

Product propensity models require minimum 18 months of product acquisition and closure history with customer features at time of event
Pricing optimization should be rolled out incrementally (one product at a time) to avoid unintended competitive responses
Competitive rate monitoring requires either automated web scraping of public rate pages or subscription to a competitive intelligence provider
Campaign attribution requires Salesforce FSC integration with proper UTM tracking for digital channels and advisor activity logging for branch channels
Bundle design should involve product managers from each product line to validate commercial viability and risk appetite

← Back to Catalogue | Previous: Loan Lifecycle | Next: Operational Risk →