Product
Lost pallet search, without the search.
A misplaced pallet is not lost. It is somewhere else, and somebody still has to find it. The difference is that the answer is already sitting in a recording nobody has watched yet, so a search becomes a query instead of a walk.
What is actually missing
Nobody loses a pallet. They lose track of it.
The pallet is on a shelf somewhere in the building. It has a location, a barcode and a weight. What is actually missing is the record connecting that pallet to that shelf, and a manual search rebuilds the record by walking the racking one aisle at a time until somebody's eyes land on the label.
That is why the search is expensive and why it is expensive every single time. A picker who cannot find a pallet at its recorded location stops, escalates, and a second person joins the hunt. Multiply that across a 30,000 m² site over a year and the labour adds up to €100,000 to €200,000, on top of whatever the audit itself costs. It is rarely booked under "stocktaking" in the budget, which is one reason it survives unexamined for years. The full breakdown of where that money goes sits on the cost of manual stocktaking.
The mechanism
A search becomes a query, because the drive already answered it.
Every run down an aisle registers every location it passed, occupied or empty, at up to 3,000 units per hour. The pallet you are looking for was almost certainly captured on a recording that already happened.
- The location is already known. Location level recognition reads where every pallet actually sits, not just where the WMS expects it, so a pallet that moved without a transaction still shows up in the record of the aisle it is standing in.
- You are not sending anyone to look. Instead of a person walking the racking, someone opens the discrepancy list, filters to the pallet identifier, and reads the location the last recording found it in. The recording behind that line is the evidence, not a guess.
- The next scheduled run confirms it. Because a run is a 20 minute drive covering 500 to 1,000 pallets, driven during a break, the site does not wait for an annual event to catch a pallet that wandered. See continuous cycle counting.
- The search cost falls because the labour changes shape. Nobody is paid to walk and look. One person checks a screen instead, which is where the tenfold difference at the top of this page comes from.
The search was never really about finding the pallet. It was about finding the record of where the pallet already was.
What the portal reports
Two discrepancy types cover most searches.
Physically here, recorded elsewhere.
The pallet exists and the run found it, just not where the WMS says it should be. The discrepancy line carries the actual location and the recording that confirms it, so a picker is redirected instead of dispatched to search.
A targeted question rather than an open ended hunt.
The WMS expects a pallet at a position and the run did not find it there. That narrows the search to the last recording that did show it, rather than the whole warehouse, which is the difference between a lead and a write off.
Both lines open the recording that produced them, so a disputed result is settled by looking at the image rather than sending someone back into the racking. The other two discrepancy types, and how the location data behind all four is captured, are on empty and occupied locations.
Where this shows up
95% less missing pallet labour, at scale.
A global 3PL running comprehensive audits of 30,000 to 100,000 pallets for retail and fashion clients now scans 100,000 pallets a month. Missing pallet related labour is down 95%, alongside stock accuracy of 99.97%. The full story is in the 3PL case study. Every site's own accuracy and break even figures, including how break even is calculated, are published on the benchmark page.
Questions
What managers ask about finding a pallet.
How does this actually find a specific pallet faster?
Because the recording already exists. Every drive down an aisle registers what it saw at that position, so finding one pallet is a lookup against recordings you already have, rather than a fresh walk through the racking to create a new one.
What if the pallet genuinely is not in the building?
Then no recording will show it, and that absence is itself useful. An "expected pallet missing in recording" line with no sighting in recent runs is a much stronger basis for a write off decision than a search that simply gave up.
Do we need to wait for a full stocktake to search?
No. A 20 minute run covers 500 to 1,000 pallets and can be scheduled whenever it is needed, including as a targeted zone recount, so a search does not wait for the next scheduled count.
Does this replace our WMS?
No, it reconciles against it. Sentispec Inventory works out of the box on a stock file exchange, with a standard API when you want corrections to flow back automatically. See integrations.
Next step
Stop paying people to walk and look.
A 30 minute demo showing how a discrepancy list turns a specific pallet into a location, using recordings from a site with your racking and label format.