The complaint reached the operations desk on a Tuesday: two of the warehouse’s mobile robots were idle again, and a purchase request for a third unit was already waiting for approval. The shift supervisor told the story the way operators tell these things, with a shrug and a sigh. “We keep buying parts and still cannot keep the fleet moving.” This diagnosis walks through what happened when someone suggested a different first step: a plain robot fleet inventory, counting what the site already owned before ordering anything.
The symptom
Idle machines are easy to see and hard to explain. In this case the pattern was irregular. A robot would stop for a day, return to service, then stop again. Each time, the team ordered something: a spare battery, a replacement sensor cover, a charging adapter. The storeroom shelves filled up, the budget line for consumables grew, and the idle count did not fall. Everybody agreed the fleet was short of something. Nobody could say what.
The first hypothesis, and why it failed
The obvious theory was a supply shortage. If robots stop because parts are missing, then buying parts fixes the problem. But the theory made a prediction that could be tested: if parts were truly short, the storeroom would be bare when a robot went down. Nobody had checked. The supervisor asked the storekeeper to count what was physically on the shelves, matched against what was recorded in the ordering sheet. That count was the beginning of the real work.
The inventory, step by step
The team set aside a morning and split the site into four zones: the storeroom, the charging area, the maintenance bench and the fleet floor itself. Two people walked each zone with a clipboard. They wrote down every robot, battery, charger, tool, spare and consumable, together with its condition and where it was found. They did not fix or move anything during the count. The rule was record first, decide later.
- Fleet floor: which robots were running, parked or stopped, and who had last touched them.
- Charging area: chargers present, chargers working, chargers in use, and spares sitting unused.
- Maintenance bench: items awaiting repair, with the date they arrived on the bench.
- Storeroom: spares by type, quantity and condition, including anything still in its original packaging.
What the count revealed
The storeroom held more than anyone expected. There were duplicate batches of the same spare, ordered by different shifts who had not known about each other’s requests. Three working chargers sat boxed in a corner, while one corner of the charging area was short of power points. Two units waiting on the bench had been marked for repair so long that the paperwork had disappeared, which meant nobody was chasing them. And one of the “idle” robots turned out to be fully serviceable; it had been parked behind the maintenance bench and forgotten after a routine check.
The deeper finding was that the shortage was not of parts. It was of visibility. Items existed but could not be found, repairs existed but were not tracked, and capacity existed but was not in the right place.
The revised diagnosis
With the numbers on the table, the team rewrote its problem statement. Instead of “we need more spares and a third robot”, it became “we cannot match the equipment we hold to the places where it is needed, and repair jobs are slipping through the gaps”. That is a process problem with process solutions, and none of them required a purchase order.
- Relocate what already works. Move the boxed chargers into the zone short of power, and park serviceable robots where operators expect to find them.
- Give repairs a ticket. Every item on the bench gets a date, a reason and a named person.
- Merge the ordering sheet with the storeroom count. One record, updated whenever something enters or leaves.
- Hold one weekly stock look. Fifteen minutes, one person, a simple list.
How the fix held up
Within a few weeks the supervisor reported that the stop-and-start pattern had mostly disappeared. The robot that had been forgotten behind the bench went back on shift. The two long-waiting repairs were finished once someone owned them. The purchase request for a third unit was withdrawn, not because the idea was wrong, but because the team could now see its actual spare capacity and decided to revisit the question later with real numbers.
Where this fits in good fleet habits
None of this is exotic. Good operations writing on this site covers the same territory from different directions, such as the basics set out in a beginner-oriented set of fleet management habits and the routines described in an introduction to maintenance procedures for newcomers. A stock count belongs alongside both. It supplies the raw facts that scheduling and maintenance depend on, and it is almost always cheaper than discovering the same facts through failures.
Notes for anyone repeating the exercise
- Use a fixed template so every zone is recorded the same way.
- Take photographs of shelves and benches; they settle later arguments.
- Count condition as well as quantity. A spare that does not work is a hole in the list.
- Ask each shift what they ordered recently and why; the answers explain the duplicates.
- Resist fixing things during the count, or the numbers will lose meaning.
A word about this site
If you are new here, the Robots Ops home page collects the latest articles about running robotics operations, and a stock-take like this one is a good first topic for any team to try before reading further. The habit is small and the payback arrives quickly.
What to take away
When equipment keeps failing to appear where it is needed, the instinct is to buy more. The better first move is a morning with a clipboard. Count the fleet, the chargers, the spares and the repairs waiting on the bench. Separate what is missing from what is merely lost or misplaced. Then change the process before you change the budget. The supervisor in this story still keeps a copy of that first count pinned near the storeroom door, as a reminder that the answer to “what do we need?” often begins with “what do we have?”