Barcodes on the floor: scan the work instead of re-keying it
How to find the manual steps on your floor that a barcode scanner actually pays to replace, what to put a barcode on — items, locations, and lots — and how one scan captures the lot and steers the pick to the right one instead of leaving a person to key it and get it wrong.
Chef Diego runs a real food plant. If this page didn't get you there, tell us — a person reads every message.
After this lesson you can find the manual steps on your floor that generate the
most errors, decide where a barcode scanner actually pays to replace a keyboard
and a clipboard, and know what to put a barcode on so one scan captures the item,
the location, and the lot — instead of a person keying three numbers and getting
one of them wrong.
The keyed number is where the errors start
At low volume, keying numbers by hand works. Someone counts what came in, writes
the lot on a sheet, types the total into a spreadsheet later. A few orders a day,
and it holds together.
Volume breaks it. More receipts, more picks, more lots in motion, and more people
touching the keyboard — and the hand-entry starts to leak. A transposed lot code.
A case count off by two. A pick logged against the wrong lot because two looked
alike on the shelf. None of these feels like a disaster on the day it happens. The
problem is that a keyed error does not stay put — it compounds.
Trace one through the operation:
A wrong lot keyed at receiving flows into every batch that draws on it. Months
later a recall trace follows that lot back to the wrong supplier, and the real
one keeps shipping.
A miscount makes your on-hand disagree with the shelf. Reorder fires on a number
that is not real, so you run short on a material you thought was covered or
double-buy one you already had.
A pick keyed against the wrong lot means your record says one lot shipped to a
customer while a different lot actually left the door. The genealogy is clean and
wrong.
Every one of these starts as a single keystroke. A barcode replaces the keystroke
with a read: the scanner pulls the exact item, location, or lot off a printed code
instead of asking a tired person to type it correctly at the end of a shift.
Where scanning pays
You do not scan for its own sake. Scanning earns its place at the handful of
moments where a person would otherwise copy a number from one place to another —
and getting that number wrong costs you downstream. Four of those moments carry
most of the value.
1
Receiving
As goods come off the truck, scan the item and capture the lot and expiry from
the supplier's label — or print your own lot label and scan that — against the
purchase order. The count and the lot land in the record at the dock, where the
three-way match
and later traceability both need them, instead of being written on a sheet and
keyed in an hour later.
2
Picking
When stock is pulled — for a batch or an order — scan the lot as it leaves the
shelf. The record then matches what physically moved, and the scan is the moment
you can check the pick against a rule rather than trusting whoever reached in.
3
Production
Scan each ingredient lot into the batch as it is consumed, and scan the finished
lot as it comes out. That builds the link from inputs to output as the work
happens, with no separate clipboard step to reconcile later.
4
Shipping
Scan the finished lot onto the order or pallet as it ships. Now you know exactly
which lots went to which customer — the one-step-forward record any trace needs —
without reconstructing it from memory after the truck is gone.
None of this is a regulation. It is standard operations practice: put the capture
at the point of the action, and let a scan do the copying a human would otherwise
do by hand.
What you put a barcode on
A barcode is just a machine-readable label for one thing. On the floor, three
different things are worth labeling, and they answer three different questions:
The item — what is this? The material or product, the same across every
unit of it.
The location — where is it? A bin, a rack, a cooler, a dock. Scan a
location and the system knows where stock sits.
The lot — which batch is this, exactly? This is the one that carries
traceability, and it usually earns its own label: a
you stick on the pallet or case when it is received or
made.
A scan of a location plus a lot says "this exact lot is here" in one motion. That
combination is what lets a scan capture everything a person would otherwise key —
and it is why the lot label, done at
receiving or at the moment you make the lot,
is the one to get right first.
The scan can steer the pick, not just record it
Capturing the lot is half of what a scan is good for. The other half is using the
scan to enforce a rule at the moment of the pick — and the rule that matters most
on perishables is . As the
FEFO lesson
puts it, a rotation rule only works if it reaches the person with their hands on
the stock. A scan is where it can. When the picker scans a later-expiring lot while
a sooner-expiring one is still available, the scan can flag it right there, before
the wrong lot leaves — FEFO enforced at the pick, not audited after it.
The same scan writes the lot into the record as the work happens. Scan the input
lots into a batch and the finished lot out, and you have built the
that the
one-up-one-down lesson
describes — without re-keying it from a clipboard afterward. Software built for
food can do this: capture the lot straight from the scan at receiving and
production, so the genealogy is written as the work happens rather than
reconstructed later. Bettr Manager, the platform this site is part of, works that
way. It is one way to get scan-capture; a disciplined habit of scanning printed lot
labels and logging them by hand reaches the same record with more work. For one
concrete way to book a receipt in against its PO and capture its lot at the dock,
the docs walkthrough on
receiving a purchase order shows the
flow.
Start with your worst step, and measure it
You do not barcode the whole floor at once. Find the single manual step that
generates the most errors — usually receiving or picking — and pilot scanning
there, capturing the lot and the location. Then measure whether it worked.
Count the errors before and after
Pick a number you can actually count: receiving corrections per week, mis-picks
caught, or times the on-hand had to be fixed by hand. Write down the rate for a
couple of weeks before the scanner goes in, and the same number after. An
illustrative before-and-after — corrections dropping from several a week to nearly
none — is the kind of evidence that tells you whether to extend scanning to the
next step or fix the pilot first. Use your own counts, never a vendor's promise.
A scanner only helps if the crew actually uses it, which means the scan has to be
faster than the clipboard it replaces, not slower. That is an adoption problem, not
a hardware one, and
getting the floor to use the system
covers it directly.
Where this leads
Barcodes cut the errors that come from re-keying inside one location. The record
matches the shelf, the lot is captured where the work happens, and the pick goes
out in the right order. As you add a second site or a second legal entity, the same
lots start moving between them, and the question changes: how do you keep one
source of truth when stock — and the lots that carry its traceability — crosses
from one place to another? That is the next lesson.
