PM Quest · Practice cases · Learning resources

Stripe Canada SMB Transaction Volume Drop Investigation

Root Cause Analysis · beginner

As a Product Manager for Stripe's SMB payments platform, you've just been alerted to a sudden, 18% drop in successful transaction volume from Canadian merchants over the past 24 hours. This is a critical growth segment. How would you approach investigating this unexpected decline?

Write your approach on PM Quest

Community approaches

Alfiya’s approach

  1. Metric Validation & Clarification Before alerting engineering teams, the signal must be verified to make sure it is not a reporting glitch:
  • Data Pipeline & Telemetry: Database logs and ingestion pipelines were checked for latency. Reporting dashboards are updating in real time with zero ingestion lag.
  • Seasonality & Calendar: The drop was measured week-over-week rather than day-over-day to account for normal daily cycles. There are no Canadian statutory holidays on May 15.
  • Confirmation: The 18% drop is genuine. Successful transactions have actually dropped across Canadian merchants.
  1. Segmentation & Narrowing Down To find where the failure lives, the data was sliced across multiple dimensions:
  • Client & Platform: The decline is evenly split across iOS, Android, and web browsers. It affects multiple app and API versions, ruling out a single front-end release issue.
  • Merchant Profiles: Both small and medium businesses are impacted uniformly across all Canadian provinces. New and existing merchant accounts show the same drop.
  • Payment Rails & Networks:
    • Transactions on global credit card networks (Visa, Mastercard, Amex) are processing normally with healthy conversion rates.
    • Transactions over Interac debit rails plunged by roughly 90%.
  • Transaction Logs: Total checkout attempts remain steady, but Interac attempts are failing with a standard error code: gateway_timeout (HTTP 504).
  1. Hypothesis Generation & Technical Diagnosis With the failure isolated to Interac debit rails timing out, we divide hypotheses into internal and external checks:
  • Internal System Checks (Stripe):
    • Checked if the team pushed any recent code deployments or service changes to Canadian rails in the last 48 hours. (Result: No deployments or configuration changes occurred.)
    • Checked if internal API timeout thresholds, signing keys, or partner SSL certificates expired. (Result: All internal configs, tokens, and certificates are fully valid.)
  • External Dependency Checks (Interac & Issuing Banks):
    • Checked Interac's public status page. It reported operational status, but outbound requests from Stripe to Interac's API endpoints showed sustained latencies exceeding 15,000 ms before timing out.
    • Sliced the timeouts by customer issuing bank. This revealed that 85% of all failed Interac requests came from accounts held with a single institution: TD Bank.
  • Root Cause: TD Bank's internal gateway on the Interac rail became unresponsive, causing requests to hit our 15-second timeout limit and dropping Canadian SMB transactions by 18%.
  1. Immediate Mitigation & Fallback Strategy Immediate mitigation focuses on keeping shoppers from failing checkouts and saving merchant sales:
  • Trigger Circuit Breaker: Temporarily suppress/disable Interac routing specifically for TD Bank accounts so shoppers do not get stuck on a 15-second loading spinner that ends in a failed purchase.
  • Smart UI Fallback:
    • When a customer picks Interac and selects TD Bank, display a clear inline notice: "TD Bank Interac is currently experiencing delays. Would you like to use a Debit/Credit card or Apple Pay instead?"
    • Reprioritize working payment options (Visa Debit, Mastercard Debit, Credit Cards, Apple Pay, Google Pay) to the top of Stripe Checkout to keep conversion high.
  • Partner Escalation: Contact technical points of contact at both Interac and TD Bank, sharing error traces, status codes, and latency spikes so their infrastructure team can resolve their server failure.

N’s approach

metric in focus : dip in successful transaction for over 24 hrs Issue tree : System issue : any changes in the metrics calculation? internal factors : any new feature development, any changes in the regular flow, product cannibalism external factors : change in market trends? a other factors : any change in government regulation for fintech?