At a glance
Role and impact
Project Leader (Product Owner). Over this period the team grew from 4 to 7 people.
5% stockout reduction across the rollout, and less overstock.
Methods and tools
Recommendations vs sales analysisLead time in the forecastCoordination with the ML teamBackend APIsRollout in 5 countries
Background
The team owned Restrictions, a seller-inbound control process that ran from a spreadsheet. We replaced it with backend APIs that automated the checks and the seller messages, and rolled it out in Argentina, Brazil, Mexico, Chile and Colombia. Recommendations was the next step: from blocking what shouldn't come in, to advising what should.
What I did
- Led the team's analysis linking recommendations to actual sales.
- We found the sales forecast behind them ignored inbound lead time, so it always came up short: it didn't cover the sales made while the units were on their way to the warehouse.
- Drove a new calculation that included it, coordinated with the Machine Learning team.
Result
- Recommendations vs sales
- →
- Lead time missing
- →
- New calculation
- →
- −5% stockouts
5% stockout reduction across the rollout, and less overstock.