From quarterly waterfalls to a release every three minutes - 1,800 stores.
The retailer's digital and store technology teams were moving at completely different speeds. We rebuilt the release model around product squads, feature flags, and shared telemetry so store operations could trust rapid change instead of fearing it.
The situation we walked into.
The retailer had modern commerce ambitions, but store systems still depended on slow weekend release windows and manual certification. That meant even simple checkout or loyalty changes dragged through governance queues for weeks.
The underlying issue was organisational. Product, store operations, payments, and merchandising all had different release calendars and different tolerance for risk. We had to make rapid delivery feel safer than slow delivery.
A simplified view of the delivery shape, the control points that mattered, and the signals the client team used to keep the program on track.
The changes that made the outcome possible.
Platform engineering
A shared delivery platform standardised build pipelines, secrets, progressive delivery, and rollback automation.
Store-safe releases
Feature flags, dark launches, and cohort routing let the client test store changes without nationwide exposure.
Experience telemetry
Digital and in-store journeys were monitored on the same service map so defects could be isolated quickly.
Operating model reset
Quarterly program boards were replaced with weekly product reviews and release readiness scorecards.
How the delivery moved from pilot to scaled operation.
Value-stream mapping
We identified release bottlenecks across POS, e-commerce, payments, and fulfilment.
Golden paths
The first product teams moved to standard pipelines, automated testing, and on-call accountability.
Store rollout
Pilot stores adopted progressive releases with playbooks for trade days, peaks, and local rollback.
Enterprise cadence
Merchandising and operations leaders shifted planning into a continuous prioritisation rhythm.
What changed after the transformation settled into the run.
Trading became more responsive
Product teams now launch promotions, fulfilment tweaks, and checkout changes around market signals instead of around release windows.
Store confidence improved
Regional managers gained live dashboards and clear release notices so they knew what changed and how to escalate issues.
Engineering quality improved alongside speed
Teams measured lead time, failure rate, and recovery together, which cut heroics and improved planning accuracy.
We thought we had a speed problem. What we actually had was a trust problem between digital, stores, and operations.Chief Technology & Transformation Officer - Global retailer