Home/Industries/Retail & Inventory
IndustryStock accuracy first. Everything else follows it.
If the system says an item is in the store and it is not, every other retail technology built on top of that record is compromised.
What is actually hard here.
Inventory records drift. Items are mis-scanned, moved, returned to the wrong place or taken. The record and the shelf diverge quietly, and every process downstream — replenishment, online availability, click-and-collect — inherits the error.
A manual count is accurate on the day and stale within a week, which is why it is done rarely, which is why the drift compounds.
What it looks like on site.
The outputs worth having.
- Stock accuracy — what the system says is present against what a bulk read finds.
- Shelf availability — the item exists in the building but is not on the shelf, which is a different problem from being out of stock.
- Shrinkage patterns — where and when, rather than a total discovered at year end.
- Replenishment triggers — from counted reality rather than from assumed sell-through.
- Footfall and dwell by zone — from radar, anonymous by construction.
- Conversion — people in against transactions out, per zone, per hour.
Why these, and not the others.
RFID leads, and it is not close. Item-level tags cost a small fraction of anything powered, and a handheld or an overhead sweep reads hundreds at a time — which turns stocktaking from an event into a routine. This is the mature, proven application of the technology.
mmWave supports for people-counting. It gives footfall and dwell without a camera, which sidesteps the customer-privacy question entirely rather than managing it.
UWB is usually unnecessary here. No decision a shop makes changes because a jumper's position is known to the nearest handspan.
A sensible order of work.
- Tag at source if you canTags applied by the supplier are far cheaper than tags applied in a back room.
- Count before you automateEstablish the current accuracy figure. Without it there is nothing to prove.
- Handheld before fixedA handheld sweep proves the value at low cost. Fixed overhead readers come after the case is made.
- Portals at goods-in and the doorReceiving accuracy and loss prevention, both from the same infrastructure.
- Radar at entrancesFootfall that nobody has to consent to being filmed for.
If it works.
- Stock accuracy becomes a number that is measured weekly instead of guessed annually.
- Online availability can be trusted against store stock.
- Footfall data exists without a camera anywhere in the building.
Does this look like your site?
Tell us the space, what moves in it, and the decision a position would change. We will tell you which technology fits and what a first deployment would involve.