A transportation company quotes you a transit time. It starts when the truck is loaded. The customer's clock starts much earlier, when the order was confirmed.
Everything in between belongs to you, and almost nobody measures it. This article shows how to measure it for one order in an afternoon, how to read the result, and why fixing the biggest wait is often not the fix that helps.
Where your clock starts and the carrier's starts
Thailand's Ministry of Industry has an indicator set for factory logistics that splits the order into stages on purpose. The order cycle runs from the company confirming the order to the customer receiving the goods. Inside it, one stage covers passing the order to every department that needs it. Another, the delivery cycle, starts only when the goods are loaded on the vehicle.
That split is the point. The carrier is responsible for the last stage. The first stages are yours, and a carrier who is fast on the road cannot get back hours that were spent before the truck arrived.
Notice also that the indicator set counts in days. A dispatch desk that works to a same-day cut-off needs hours, and in this case needs the hours broken down by step.
Map one order, with real timestamps
The textbook's method is plain. Walk the actual process, write down each step and how long it took, mark which time is work and which is waiting, work out the shares, then look for causes.
The steps for a typical warehouse order are: order confirmed, passed to the warehouse, stock and credit checked, waiting to be released for picking, picked, consolidated and staged at the dock, waiting for the truck, truck waiting to load, loaded.
Here is one order, invented for the illustration. Every row is a wait unless someone is touching the order.
| Step | From | To | Elapsed (min) | Someone working (min) |
|---|---|---|---|---|
| Order passed to warehouse | 10:05 | 11:40 | 95 | 10 |
| Stock and credit check | 11:40 | 14:00 | 140 | 15 |
| Waiting for the pick wave | 14:00 | 15:30 | 90 | 0 |
| Picking | 15:30 | 17:10 | 100 | 70 |
| Consolidate and stage | 17:10 | 18:00 | 50 | 35 |
| Staged load waits for the truck | 18:00 | 07:30 next day | 810 | 0 |
| Truck at the gate until loading starts | 07:30 | 08:20 | 50 | 0 |
| Loading | 08:20 | 09:40 | 80 | 60 |
| Total | 10:05 | 09:40 next day | 1,415 (23 h 35 min) | 190 (3 h 10 min) |
Someone was working on this order for 190 of 1,415 minutes, or 13%. The other 87% was waiting. Your own figure will differ, but you will not know it until you map it, because nobody can see it from inside any single department.
The biggest wait is not always the one to fix
The longest block is the 810 minutes the staged load sat overnight, 57% of the whole. It is also the least useful place to start. The truck was booked for 07:30, so the order was never going to leave earlier.
Now look at the waits before picking: 95, 140 and 90 minutes, 325 minutes in all. The load was ready at 18:00. With those 325 minutes cut by 240, it would have been ready at 14:00. That does nothing for a truck booked for 07:30. It does two other things:
- A later cut-off. If an order takes four fewer hours to become a loaded-ready pallet, the time by which a customer must order to get next-morning loading can move four hours later. That is a service you can sell, and the order rules that set transport cost are where it is written down.
- A same-day truck becomes possible. A load ready at 14:00 can be collected that afternoon, if you gave the carrier notice and the driver has hours left.
So read the map in two parts. Time before the load is ready decides what you can promise. Time after it is ready decides what you pay for, because that is where trucks stand. The 50 minutes at the gate is the first, smallest piece of waiting at your own dock.
Find the cause, not just the step
Once you know which step to attack, ask why it takes that long. The textbook groups causes under four headings, which is enough to start with:
| Heading | Examples that show up in a dispatch map |
|---|---|
| People | Only one person can release holds. Shifts end before the work does. Nobody owns the order between two departments |
| Place | The staging area is full, so finished orders wait for space. Long walks between pick and dock |
| Equipment | One forklift for picking and loading. Paper pick lists. No scanner at the dock |
| Method | Picking released in waves rather than as orders arrive. Credit check done after stock check. Rush orders jumping the queue and delaying everything behind them |
In the illustration, the two waits that can be shortened are method. The 90-minute pick wave is a rule about when work is released, not a shortage of people. The 140-minute check is stock and credit checked one after the other when they could be checked together. Changing a rule costs little. Equipment costs money, so look at method first.
What to do on Monday
- Choose 20 orders from the last two weeks. Include some that arrived late in the day, because those usually behave differently.
- For each, find six timestamps: order confirmed, order reached the warehouse, picking started, load staged, truck arrived, truck left. Your system, the gate log or the driver's delivery note usually has most of them. Fill gaps by asking, not guessing.
- Work out the elapsed time for each step on each order. Take the middle value for each step, and write down the longest.
- Add up the time someone was working, even roughly. If it is a small share of the total, as it is in the table, you have a waiting problem, not a speed problem. Hiring more pickers will not fix it.
- Split the map in two at the moment the load was ready. Fix the first half to move your cut-off. Fix the second half, with the carrier, to cut standing time.
- Run the map again a month later with the same method. The comparison is the proof.
The measure that decides whether the work paid off is on-time delivery counted from the promise you made, because that is the clock the customer is holding.
Where the evidence stops
No Thai study of hours between order confirmation and truck departure was found. The method comes from the textbook, the stage definitions come from the Ministry of Industry's indicator handbook, which counts in days, and every time in the table is an illustration. Your own twenty orders are the real data, and they will not look like this table.