Investigate Checkout Conversion Spike

Business Context

ShopWave is an e-commerce marketplace for home goods. Yesterday, the executive dashboard showed a sudden spike in checkout conversion rate, and the VP of Growth wants to know whether this reflects real performance improvement, a mix shift, or a tracking issue.

Metric Scenario

For the prior 4 weeks, checkout conversion rate was stable at 3.8% to 4.1% of sessions. On Tuesday, it jumped to 6.2%. At the same time, daily sessions fell from 1.25M to 910K, orders increased only slightly from 49K to 56K, paid traffic share dropped from 42% to 25%, and iOS app traffic rose from 28% to 44% of sessions. A new checkout UI launched to 15% of users on Monday evening, and the data engineering team also migrated part of the sessionization logic that same night. Leadership needs a same-day diagnosis before reporting results externally.

Requirements

1. Define the KPI precisely, including numerator, denominator, and edge cases.
2. Outline a structured investigation plan to determine whether the spike is real or caused by instrumentation, denominator changes, or traffic mix.
3. Decompose the KPI into the most relevant dimensions and identify which cuts you would check first.
4. List 3-5 hypotheses for the spike and explain how you would validate each.
5. Recommend immediate next steps, including guardrails and communication to stakeholders.

Data Available

• web_sessions: session_id, user_id, session_start, platform, device_type, traffic_source, landing_page
• checkout_events: session_id, cart_view, checkout_start, payment_submit, order_complete, timestamp
• orders: order_id, user_id, session_id, order_value, discount_amount, payment_method, timestamp
• experiment_assignments: user_id, experiment_name, variant, assignment_time
• event_quality_logs: event_name, platform, error_rate, duplicate_rate, missing_event_rate, deploy_version