Most container tracking rollouts do not fail in month three, when someone finally admits the data has drifted. They fail in the first afternoon, quietly, and everybody carries on for six weeks before noticing.
I build inventory and asset tracking software, which means I get to watch a lot of first days. Across 118 small operations that started using ours, 57 recorded something real within the first hour. Fifty-eight never recorded anything at all. Three landed anywhere in between.
Three. That is the number worth staring at. Not the split, which will be different at your yard and different again next quarter, but the emptiness of the middle. Almost nobody starts slowly and gets there eventually. Whatever happens in the first hour is what happens.
There is a version of a rollout that sounds disciplined and is fatal. You buy the system, you set it up, and you agree that rather than stopping work to enter four hundred containers, the yard will add each box the first time it is touched. Natural. No downtime. It fills itself. It does not. Six weeks later there are 19 containers in the system, and nobody can tell you which ones.
The reason is worth understanding because it is not laziness. Entering a box that is not in the system yet is a different task from scanning one that is. The scan is two seconds. The first entry is a number, a size, a condition, a decision about what to call the site it is going to, and it happens at the exact moment the driver is trying to leave. So, the box gets moved and not entered, and the next driver finds the same box still missing and does the same arithmetic. Every one of them is right, individually.
A half-populated system fails the only test that matters. Somebody asks where box 214 is, the system does not know, and the answer they learn is not “box 214 was never entered.” The answer they learn is “the system does not know where things are.” One of those is a data gap. The other is a verdict, and it is the one that sticks.
The instinct is to start with the busiest boxes. That is half right and it is the half that gets people into trouble, because the busiest boxes are the ones already in motion, and a container on a jobsite is the hardest thing to enter accurately.
Start with what is standing in the yard on the day you begin. Whatever is on the ground in front of you, in one sitting, entered properly. It is a bounded job with a visible end, which is what makes it get finished, and everything in it is verifiable because you are looking at it.
Then take the boxes that come home over the following fortnight and enter each one as it lands, at the yard, not on the road. That is a second deliberate pass with a definite finish, and it is a different thing from populating as you go, even though it looks similar. The difference is that somebody owns it, and somebody knows when it is done. If your fleet is too large for one sitting, split by yard and finish one yard completely. A finished yard is a working system. A part-finished yard is a permanent excuse.
Name one person and give them the afternoon. This is the part that gets delegated into failure. Spreading first entry across a crew feels efficient and produces four naming conventions, three ideas about what “damaged” means, and nobody who can tell you what is left to do. The work is not hard. It is only unforgiving about consistency, which is a property of one person and not of a team.
The right person is usually whoever already knows where the boxes are, and that is rarely the person who signed the invoice. It is the dispatcher or the yard foreman. Once the initial set is entered, scanning belongs to whoever moves the box, which is the driver, and that split matters: entry is a project with an owner, scanning is a habit with a crew.
One decision that looks small in week one and is difficult to reverse. The instinct is to label the yard too. Bay four gets a placard, the wash pad gets a placard. It feels thorough, and it doubles the work at every stop, because now the driver scans the box and then hunts for the placard. The second scan is the one that gets skipped when it is raining.
Tag the container. Let the location be something the driver picks from a short list. The box has an identity worth defending鈥攊t has a number, it is worth real money, and it is the thing you are trying to find. A parking spot is a fact about today.
Somebody named, one afternoon, every container currently in the yard entered properly, labels on the boxes rather than the ground, and a second pass booked for the ones that come home. None of that is about the technology, which is the uncomfortable part. The tags work, the phones work, the software mostly works. What decides it is whether a real entry happened on the first afternoon and whether anybody could say when it was finished.
One more thing, because it is where the good rollouts still die: if nothing in the first fortnight tells the yard something it did not already know, people stop opening it. A system that only repeats what you fed it is a receipt, not a tool. Decide in advance what question you expect to be able to ask by then and check that the answer exists. Three out of 118 got there gradually. Plan for the hour you actually get.