Skip to content

INSIGHTS.md

Latest commit

 

History

History
108 lines (67 loc) · 7.02 KB

INSIGHTS.md

File metadata and controls

108 lines (67 loc) · 7.02 KB

CardPulse Analytics - Business Insights

Card payments analytics for the German retail banking market, covering 250,000 customers and 1.5 million transactions over 2023-2025.

Executive Summary

Five findings a head of cards and payments would want to know first:

  1. Revenue grew 48% over 3 years (EUR 42.3M in 2023 to EUR 62.6M in 2025) with steady year-over-year growth of ~21%.

  2. 23% of customers (Champions tier) drive 63% of revenue. Customer retention strategy must protect this group above all else.

  3. Premium credit cards generate 58x more profit per card than Standard debit (EUR 394 vs EUR 6.71). Acquisition strategy should shift toward Premium upgrades.

  4. 7,400 high-value customers (EUR 7.7M in revenue) are at risk of churning. A targeted win-back campaign could recover an estimated EUR 1.5M.

  5. Christmas seasonality drives a +21% December revenue spike every year, followed by a -13% January dip. Operational planning should anticipate both.

Detailed Findings

Revenue and Growth

  • Total revenue across 3 years: EUR 156.7M from 1.47M completed transactions.
  • Year-over-year growth is consistent: +22.29% (2024) and +20.89% (2025). The 1.4 percentage-point dip is a natural "larger base" effect rather than a real slowdown - banks growing 22% then 21% are considered remarkably stable.
  • December consistently spikes +21% versus November (Christmas effect). January then drops -13% as post-holiday spending slows.
  • 2025 December peaked at EUR 7.0M versus EUR 4.8M in December 2023 - a 45% growth in peak month revenue over 2 years.

Customer Segmentation

  • Premium customers (5% of base) generate EUR 3,663 in revenue each per year. Low-segment customers generate EUR 173 each - a 21x gap.
  • High + Premium combined = 25% of customers but 69% of revenue. Classic Pareto concentration.
  • Low segment activation rate is 57% versus 97% for Premium. Nearly half of Low-segment customers acquired never transact.
  • Customer behavior is geographically uniform. Berlin, Hamburg, and Munich differ in customer count, not in per-customer spending (~EUR 625 across cities).

Card Portfolio Profitability

  • World Elite (Premium tier) generates EUR 394 profit per active card through annual fees, interchange, minus rewards payouts.
  • Classic Debit generates EUR 6.71 profit per active card - 58x less profitable than World Elite.
  • Premium credit cards have 94-97% activation rates. Standard debit cards sit at 77% - meaning 23% of issued Standard debit cards are dormant.
  • The 60K Premium cards (16% of card portfolio) generate 78% of card profit. Heavy concentration of profit in the top tier.

Cross-Sell and Multi-Card Customers

  • 60% of customers hold only 1 card but contribute just 21% of revenue (EUR 222 per customer).
  • Customers with 2 cards generate 3.9x more revenue (EUR 867 per customer).
  • Customers with 3 cards generate 13.6x more revenue (EUR 3,016 per customer).
  • Card count is the single strongest predictor of customer value. Converting 1-card customers to 2-card holders is the highest-ROI growth lever.

Merchant and Spending Behavior

  • Retail drives 34% of revenue on just 16% of transactions - the highest concentration ratio of any category. Customers buy fewer items but spend more.
  • ATM withdrawals account for 22% of card volume. Cash is still significant in Germany, with EUR 31M across 6 ATM networks (Deutsche Bank, Postbank, Commerzbank, Sparkasse, DKB, ING).
  • Travel transactions average EUR 456 - 8x the average grocery transaction. Strong candidate for Premium card partnerships.
  • Dining is engagement, not revenue. 20% of transactions but only 5% of revenue (EUR 26 average ticket).

Fraud and Risk

  • Total fraud loss: EUR 2.35M over 3 years at a 1.49% per-transaction rate.
  • 1 in 10 customers experienced fraud at some point. Per-customer exposure (10%) is the metric that drives churn - far more impactful than the 1.5% per-transaction rate suggests.
  • 46% of fraud incidents occur during afternoon hours (12:00-17:00). Drives fraud team staffing decisions - the 12-17 window needs the heaviest analyst coverage.
  • Only 3% of fraud incidents occur overnight (00:00-05:00). Minimum night-shift coverage needed in this dataset.
  • Retail has the highest fraud revenue loss (EUR 813K) due to higher transaction values, even though fraud RATE is uniform across categories.

Customer Lifecycle (RFM Analysis)

  • Champions tier: 47,535 customers (23% of active base) generating 63% of revenue. The highest value group.
  • At Risk High Value tier: 5,232 customers worth EUR 1,115 each. Used to be Champions, now 273 days quiet. Urgent win-back opportunity.
  • Cant Lose Them tier: 2,173 customers worth EUR 875 each, 496 days quiet. Personal outreach required.
  • Lost tier: 33,108 customers (16% of base) but only 2.55% of revenue. Stop investing retention budget here.

Recommendations

  1. Launch a Premium upgrade campaign for Mid-segment customers. With 28K Potential Loyalists generating EUR 22M today, converting even 5% to Premium would add EUR 7M in annual revenue.

  2. Implement a 90-day churn-warning system. Flag any High or Premium customer who has not transacted in 60 days. Trigger personal outreach before the 90-day "At Risk" threshold.

  3. Reallocate marketing spend from Low-segment acquisition to Premium retention. Low segment costs are not recovered (43% never activate). Premium has 97% activation and 21x customer value.

  4. Right-size the Standard Debit portfolio. With 23% dormancy across 160K cards, an activation campaign or product retirement could remove a meaningful share of operating overhead.

  5. Staff the fraud team to peak afternoon hours. With 46% of fraud incidents occurring between 12:00-17:00, fraud analyst coverage should weight heavily toward this window.

Data Limitations

Three limitations identified during analysis that would not exist in production banking data:

  1. Online channel share is flat at 14% across 2023-2025. In real banking data, online share grows ~2 percentage points per year as digital adoption increases.

  2. Fraud rate is uniform (~1.5%) across merchant categories. Real banks see 3-5x variation, with Travel and high-ticket Online Retail typically showing the highest fraud rates. In this dataset, fraud volume tracks transaction volume rather than category risk.

  3. Tenure-normalized revenue does not compound across cohorts. All transactions in the dataset fall within 2023-2025 regardless of customer signup year, so older cohorts cannot show the long-term value accumulation real banks observe.

Methodology

  • Database: PostgreSQL hosted on Supabase (Frankfurt EU)
  • Data scale: 9 relational tables, ~2.42M rows total
  • Analysis tools: SQL (18 analytical queries), Excel (5 worksheets including PivotTables, XLOOKUP, Power Query), Power BI (2 dashboards, 11 DAX measures)
  • Sample: Synthetic German banking dataset generated to match real ISO 18245 MCC codes and realistic merchant brand distributions