Supply Chain Control Tower Automation
Automated replenishment and vendor scorecards for a consumer packaged goods network spanning 87 SKUs.
Tech stack
Outcomes
- Saved 15 analyst hours per week in manual updates
- Lifted on-time-in-full delivery by 6.2%
- Triggered exception workflows 4 hours sooner on average
Problem
Mondelēz relied on weekly spreadsheets to coordinate allocations with co-manufacturers. Lead times spiked whenever weather or packaging shortages hit, yet the ops team only learned about shortages during their Monday meeting. The analysts spent Fridays reconciling four vendor trackers by hand.
Approach
I designed an automation-first control tower:
- Modeled allocation and vendor scorecard data in BigQuery so every SKU shared the same canonical metrics.
- Built n8n workflows that polled vendor APIs and EDI drops, normalizing events into the warehouse.
- Created Google Sheets extensions that surfaced live metrics with traffic-light formatting for planners.
- Implemented Slack alerts with rich context that routed incidents based on SKU criticality.
Every automation respected the human handoffs already in place. The Slack alerts deep-linked to a templated response doc, and planners could pause flows during audits. We documented every workflow in a shared runbook and embedded observability dashboards into Looker Studio.
Outcome
The control tower cut the manual reconciliation work by 15 hours each week and increased on-time-in-full delivery by 6.2%. Exception workflows fired four hours sooner, giving procurement space to resolve packaging shortages. The COO later extended the n8n architecture to other product lines because the team could reason about incidents with far less noise.