BUY-ONLINE
-PICKUP-IN-
STORE
Building the foundation first,
then scaling what works.
THE
PROBLEM
Platform migration
plus new capability, all at once
Forever 21 needed to migrate from a legacy order management system to Salesforce OMS. The company was also Salesforce's pilot customer for their new OMS product, which added complexity. On top of the migration, leadership wanted to launch buy online pick up in store (BOPIS) across all 411 stores nationwide.
The initial ask included building full automation from day one to match feature parity with the old system. I pushed back and proposed a phased roadmap instead: migrate the foundation first, launch BOPIS second, then build automations in Phase 3 based on observed usage patterns rather than theoretical needs.
Leadership bought in and we moved forward with that approach.
THE
APPROACH
PHASE 1
Migrate to Salesforce OMS
All order transaction mechanisms were ported from the legacy system to Salesforce OMS. As the pilot customer for Salesforce's new OMS product built on Service Cloud, we hit platform bugs and worked directly with Salesforce engineering to resolve them.
The goal here was stability. Get orders flowing reliably before adding new capabilities on top.
PHASE 2
Launch BOPIS Nationwide
With a stable OMS foundation, we built BOPIS. I worked with a small engineering pod to implement the backend fulfillment logic and with design to update the website so customers could select store pickup on product pages and at checkout.
The hardest part was real-time inventory visibility across channels. Customers needed to see accurate stock at their chosen store before placing the order. After testing in a few pilot stores, we rolled out to all 411 locations at once.
Dashboards were set up in Tableau to track inventory accuracy and store fulfillment health across the network.
PHASE 3
Automate Based On Learnings
Once the system was running, patterns emerged. Customer service was spending significant time on manual order status updates and exception handling. We automated those workflows and saved 1,456 hours of manual work annually.
The bigger automation came from observing orders getting stuck in limbo due to inventory sync issues between the OMS and the legacy warehouse system. I worked with engineering to build an automated correction mechanism that unstuck those orders and routed them properly. That single automation preserved over $1.2M in potential lost revenue per year.
THE
RESULTS
Live nationwide.
Measurable efficiency gains.
Revenue protected.
411
Stores Live with BOPIS
1,456
Customer Support Hours Saved Annually
$1.2M+
Revenue Preserved per Year
The project took roughly 9 months from Phase 1 kickoff to Phase 3 completion. BOPIS launched successfully across all stores with real-time inventory visibility. The automations built in Phase 3 were targeted at observed pain points rather than theoretical ones, which meant higher ROI and faster time to impact.
THE
TAKEAWAY
Build for what you see,
not what you guess
The phased approach was the right call.
Automating before understanding real system behavior wastes engineering effort on the wrong problems.
By running the MVP first, we identified the automation opportunities that actually moved the needle on revenue and operational efficiency. The result was better ROI and less throwaway work.