EN
Back to the archive

The encyclopedia · Engineering & Operations · Strategic decision · 2020-2025

Flipkart's planning platform lifted one-day deliveries 50%.

Flipkart's planning platform combined forecasting with optimization, lifting one-day deliveries 50% and scaling new lines 5x with no new hires.

Flipkart

the move

As India's e-commerce penetration jumped during the pandemic, Flipkart's manual supply-chain planning hit its limits. Different teams planned in different directions, and growth was met by hiring more planners rather than improving the system.

The company spent over three years building a central planning platform. A machine-learning forecasting layer predicts demand at various granularities, while an optimization layer decides capacity, inventory, and how shipments flow through the network. Stochastic programming runs every 15 minutes at every warehouse to allocate capacity for maximum delivery speed.

The result was about a 50% increase in one-day deliveries and stronger speeds for priority customers, plus lower inventory and manpower costs. In some new business lines, operations scaled 5x over six months without adding a single person, and planning cycles that took 10 days fell to one.

why it works

  • Manual, loosely-coupled planning could not keep pace with rapid growth
  • Forecasting is only valuable if an optimizer acts on it
  • A continuous, 15-minute optimization loop keeps decisions current
  • Interpretability reports and governance built trust with operations teams
the payoffTie forecasting and optimization into one platformclever

what transfers

Combining prediction and optimization so they run on the same cadence turns supply chain planning into a platform that scales with volume, not headcount.

what came after

Flipkart was named a 2025 Franz Edelman Award finalist. The platform cut planning cycle times from 10 days to one, let teams scale without hiring, and enabled Flipkart to compete in the emerging 10-minute delivery market.

references

spotted an error? The archive wants to know.

same kind of clever