The situation
This client needed a digital retail execution solution across more than 5 million stores, on a tight timeline, after two previous attempts with other strategic partners had already failed. That history mattered as much as the technical brief — it meant limited remaining budget for a third try, and an organization that had reasonable grounds to be skeptical that this attempt would go any differently.
Why we didn't propose one big launch
A single high-risk go-live is the fastest way to repeat whatever went wrong the first two times. With that many stores, markets, and user profiles in scope, a big-bang launch concentrates risk instead of managing it — any one failure point can take down the whole rollout at once, at exactly the moment trust is thinnest.
The approach: architecture first, then a front end that could sync at scale
We designed the backend architecture on Salesforce around the business functions that actually mattered, then built a front-end SaaS application with bi-directional sync to Salesforce — built specifically to handle large data volumes across diverse markets and user profiles. That sync layer was the technical piece that made a phased rollout possible at all: each module could go live independently without the backend and front end drifting out of sync.
What we'd do differently
The phased approach was the right call, and we'd make it again. The refinement we'd push for earlier next time is a tighter, shared definition — with the client, in writing — of what "successful module go-live" means before the first phase ships, so each phase has its own clear pass/fail bar rather than being judged only against the overall rollout.
The outcome
A phased, module-by-module rollout across markets — avoiding the single high-risk launch that had effectively sunk the two previous attempts.