Glossary
Feature Flag
A toggle that turns product features on or off without redeploying code.
By Amit Tyagi, Fitoor Capital · AletheiaAI Glossary
Definition
A feature flag is a conditional statement in code that controls whether a feature is visible or active to users. Instead of deploying new code to release a feature, engineers wrap it in an if-statement tied to a flag. Flip the flag, and the feature activates for specific users, regions, or percentages of traffic.
Feature flags decouple deployment from release. You can deploy code to production without users seeing it. This reduces risk: test new features with 5% of users, catch bugs, and roll back instantly without a new deployment.
Common patterns include canary releases (gradual rollout to detect issues early), A/B testing (compare feature variants), and kill switches (disable broken features instantly). Teams use tools like LaunchDarkly, Harness, or open-source solutions like Unleash to manage flags at scale.
Feature flags are essential for continuous deployment and safer experimentation. They enable faster iteration cycles and reduce mean time to recovery (MTTR) when issues occur.
India Context
Indian startups operating at scale—especially fintech and e-commerce firms handling millions of daily transactions—rely heavily on feature flags to meet RBI and SEBI compliance requirements. Deploying a feature flagged change allows compliance teams to verify regulatory adherence before a full rollout, reducing audit friction and avoiding penalties.
Data residency laws (India Stack rules, RBI guidelines on data localization) make safe deployment critical. A feature flag lets teams test cross-region deployments in India without exposing all users to untested infrastructure changes. This is particularly important for payment processors and insurance platforms where downtime or data mishaps invite regulatory scrutiny.
Startups using feature flags report 40–60% faster feature time-to-market and lower rollback costs—measurable gains in India's competitive landscape where rapid iteration is often the difference between survival and obsolescence.
Example
Razorpay, India's largest fintech platform, uses feature flags to test new payment methods (UPI, buy-now-pay-later) with select merchant cohorts before going live nationwide. This prevents outages affecting millions of transactions and allows the compliance and payments teams to verify regulatory requirements in production with low blast radius.
When rolling out a new settlement feature, Razorpay might enable the flag for 100 test merchants first, monitor transaction success rates and data accuracy, then expand to 10%, 50%, and finally 100% of merchants—all without a single redeployment.
Frequently Asked Questions
Apply what you've learned
See this term at work on real Indian companies.
AletheiaAI checks market narratives against the filings behind them — screener, company disclosures, and sector reports across India’s listed companies, free.