Supplier lead time watcher
A small system that records when each purchase order was raised, approved, sent, acknowledged and received, works out the real distribution of lead times per supplier and product, and flags the gradual slippage that a promised lead time hides. It never switches supplier; it produces the evidence for a conversation. Seven posts on the same system, one diagram at a time, with a cost breakdown and an engineering reference at the end.
- 01
A supplier lead time watcher on AWS for a few dollars a month
The whole system on one page — timestamps, distribution, drift — and why the supplier’s own number is the wrong input.
- 02
How lead time actually gets measured
The five timestamps and what each segment means, how partial deliveries are handled, why working days need the supplier’s calendar, and the orders that must be excluded.
- 03
Why the average is the wrong number
Why an average lead time produces stockouts by construction, how to pick a percentile, why a long consistent supplier beats a short erratic one, and how to report the spread.
- 04
How slippage gets caught
How gradual drift is detected against a baseline, why single-order alerts train people to ignore alerts, what the acknowledgement segment shows first, and the seasonal confounder.
- 05
How it turns into a conversation
Why no supplier decision is automated, what a scorecard does to a relationship, why showing internal delay first works, and the annual review this makes possible.
- 06
What the supplier lead time watcher costs
About $2 a month. One Bedrock read per purchase order is the only line that grows; the queue, the table, the mail and the storage are rounding errors. Plus the three ways the bill could surprise you.
- 07
Engineering reference: the supplier lead time watcher architecture
Same system, drawn purely for engineers. Service names, region, Lambda inventory, IAM scopes, the schemas and the exact model id.
Frequently asked questions
- What is a supplier lead time watcher?
- A small serverless system that measures observed lead times from your own purchase order and receipt records, reports them as a distribution rather than an average, and flags suppliers whose times are drifting.
- Why not just use the supplier’s quoted lead time?
- Because it is a promise rather than a measurement, and reordering against it means running out whenever they are slower than usual, which is roughly half the time.
- Why is an average the wrong number?
- Reordering on the average lead time produces a stockout on about half of orders. The number you want is a high percentile, and the post on this explains which.
- Does it include our own delays?
- It measures them separately, which is the honest part. A week sitting in an approval queue is not the supplier’s fault and should not appear in their number.
- What does it cost to run?
- A couple of dollars a month. See part six.