The operation that could only see two numbers

When I took over this operation, it measured exactly two things: whether inbound units were processed before customer SLA deadlines, and whether outbound shipments left the building by end of day. Both numbers were usually green. The building still felt like it was drowning — and nobody could say why.

The two KPIs measured the edges of the operation and nothing in between. Inbound was overproducing because "faster is better." Processing teams were working ahead of what downstream could absorb. Outbound had quietly become the constraint, and work-in-progress piled up in every corner of the building. The clearest symptom was the backlog: department backlogs swung anywhere from zero to 25 weeks of work. Zero sounds good until you realize it means a downstream team starved and standing around; 25 means buried with no way to catch up. Everyone was hitting their number. The system was failing — and nobody was wrong, they just couldn't see each other.

I built throughput targets for every major process step — not just the first and last — so each department could see what the next one could actually absorb. Visibility turned staffing from a gut call into a math problem: we rebalanced labor to the actual constraint instead of the loudest department, and started catching bottlenecks in the numbers days before they showed up in the aisles. Within three months, those targets had grown into a full KPI framework connecting daily floor metrics to company-level goals.