The workflow before
Heaven Hill allocates limited bourbon and whiskey supply across distributors and markets. In-stock, on-order and previously allocated quantities sat in separate databases and spreadsheets, so planners spent hours reconciling Oracle data by hand before they could build a plan. Departments brought conflicting numbers to the same meeting, and approvals ran over email with no audit trail.
What we built
We built an allocation planning app in R Shiny, hosted on R Connect Server, that gives planners one view of in-stock, on-order and allocated product and lets them build, change and compare allocation scenarios in minutes.
Plans move through roles: publishers create them, approvers review and sign off with comments, and viewers see dashboards. Every change is logged for audit.
Automated ETL pulls from Oracle EBS into SQL Server on an incremental schedule, and the sales part cross reference between systems is now maintained automatically instead of by hand.
Before and after
| Measure | Before | After |
|---|---|---|
| Allocation planning cycle | Baseline | 50% shorter |
| Evaluating an allocation scenario | Days of data gathering | Minutes |
| Status inquiry emails and meetings | Baseline | About 60% fewer |
| Planner adoption | Spreadsheets | 100% of planners on the app |
What's next
With allocation history in one governed place, Heaven Hill can add L2 work: demand signals and past allocation outcomes feeding a recommended allocation for planners to adjust.
