On-time delivery control tower
Strategic and tactical views of on-time delivery, from headline numbers down to the orders to fix.
Recreated on dummy data. Every name, figure and date in the recreation is invented.
Result: On-time delivery rose 18+ points in two weeks after volume grew about 4x.
Problem: Volume grew about 4x and on-time delivery fell. Missed carrier pickups, package weight errors and carrier gaps in 2 cities drove the drop, and the logistics data to see them sat in separate places.
What I built:
- A strategic view with headline numbers by day, week or month, an 8-week trend, alerts and data-health flags.
- A tactical view with open orders by shipping stage, a work queue and search by order or waybill.
- An on-time waterfall and city bars against target that show what to fix first.
- One place for all the program’s logistics data, with weak sources tagged instead of hidden.
- A button that has Claude draft the weekly summary from the numbers.
How I worked with AI: I recreated it here with Claude Code in VS Code, from my own description, on dummy data.
Decisions and trade-offs:
- I slowed merchant onboarding 50% to fix the causes of late orders. The cost was a slower ramp for those weeks.
- I tagged weak data sources instead of hiding them. The cost is a warning beside some headline numbers.
- I split strategic and tactical views. The cost is two menus to keep in step.
What changed: On-time delivery up 18+ points in two weeks.
Links: The recreation is below. Dummy data generator ยท Dummy data