Metrics question

OpenAI is considering regional data processing for API customers. How would you decide which regions to launch first? Walk through the customer demand signals, regulatory constraints, operational tradeoffs, and the success metrics you would track in the first 6 months after launch.

Practice this question out loud. An AI interviewer asks it, follows up like a real interviewer would, and scores your answer. Type or speak.

What this question tests

Whether you can sequence a regulated, infrastructure-heavy launch using real demand and regulatory signals rather than a generic global rollout plan.

Gather demand signals: enterprise sales pipeline blocked specifically on data residency, support tickets citing regional requirements, and competitor availability in the same regions.
Assess regulatory constraints per candidate region, for example EU data residency expectations versus more permissive markets, and rank by both customer demand and regulatory complexity.
Weigh operational tradeoffs: infrastructure cost and latency for each region, plus the operational burden of supporting a new compliance regime.
Prioritize the region where blocked pipeline value is highest relative to regulatory and infrastructure cost, likely the EU given typical enterprise AI demand patterns.
Set 6-month success metrics: conversion rate of previously-blocked deals, latency and uptime in the new region, and support ticket volume related to data residency after launch.
Build a repeatable playbook from the first region's launch so subsequent regions launch faster.

Uses blocked pipeline value as the primary demand signal, since that is a concrete, monetizable proxy rather than a vague popularity signal.
Explicitly ranks regions by demand-to-complexity ratio instead of launching wherever is technically easiest.
Names a 6-month metric tied to the actual business goal, converting blocked deals, not just uptime.
Plans for repeatability, since regional expansion is not a one-time project but the start of an operating pattern.

Proposing to launch everywhere at once instead of sequencing by demand and complexity.
No mention of regulatory or compliance tradeoffs specific to the region chosen.
Success metrics limited to technical uptime with no tie back to the business case (converted deals).

How would you validate blocked pipeline is really blocked on data residency and not something else?
What would you do if the highest-demand region also has the most regulatory complexity?
How would you decide when to launch a second region?