Courier, logistics and hotshot
A courier run is a service call with a shorter clock
Way Delivery runs forty cars and hotshot box trucks — bank deposit pickups from Amish businesses at Bird in Hand twice a week, then food delivery. Chain of custody, barcode scans, photographs of the bag.
It is not a different product from a plumbing shop. It is a busy shop with shorter jobs and a lot more of them.
A route is a day’s plan; a commitment is a promise
Two different things get called “the route”, and confusing them is how courier software ends up with a scheduling model that cannot express either one properly.
| The recurring commitment | The route |
|---|---|
| Twice weekly to that clinic, forever. | Today, this driver, these fourteen stops, in this order. |
| It generates work, on a cadence, inside a window. | It sequences the work that exists — including the ad-hoc jobs that simply fell in the area. |
| The same object a plumbing shop uses for a quarterly PM. Authored the same way, by the same screens. | The same thing a technician stacking three PMs in one area needs. One mechanism, two businesses. |
The bank-bag run and a quarterly maintenance round in one list, because they are one object. The number is the room left on the promise.
A recurring commitment can pin the same driver, because on a route the driver knowing the stops is the service.
Chain of custody is a workflow, not a subsystem
A photograph of the deposit in the bag. A barcode scan. A signature. Captured as steps, by the same engine a technician uses to photograph a nameplate — which is why it costs nothing to add and why the answers land in real fields you can query rather than in a note somebody has to read.
- A barcode scan is a capture type on a step, not a feature area. The raw scanned string is kept verbatim, so what the scanner actually read is still there later.
- A bag is a warehouse. The driver scans the bag they are using at the start of the run, and the pickups go into it. Two bags open at once is two visits, each stamped with its own container. That needed no new concept, which is the point.
- The van is the same idea. A driver is never bound to a vehicle — they scan whichever one they took this morning. If the truck breaks and they grab another, the run continues on the new one.
Pick the vehicle at clock-in. A bag works exactly the same way.
Where you compress the photo decides whether it is evidence
This is the most technical paragraph on this site and it is here because it is the one that cannot be fixed afterwards.
The obvious way to build a photo capture is to shrink the image on the handset, because a twelve megabyte photograph over cell service is painful. That destroys the EXIF before the server ever sees it — and “we captured the metadata” quietly becomes “the phone told us the metadata.” A client-asserted timestamp and GPS is exactly what fails under challenge, and a bag of cash deposits is exactly where somebody challenges it.
-
The raw bytes cross the wire
This is the cost, and it buys the whole thing.
-
The server reads the metadata
From the bytes the device actually produced. Not from a field the app filled in.
-
The server hashes the original
This is the step everybody skips, and it is the one that actually works.
-
Then it shrinks it and stores it
And the original is discarded or kept — that is the one knob, and it is per workflow.
And the files can live on your storage — your S3 bucket, your Azure, your Drive, your own server. Why that matters →
Courier route and rooftop quarterly are two plans in one book, each with its own movability window.