On this page
Ask a distributor what changed after they went to barcodes and the answer is usually "we scan things now". That is a description of the equipment, not of the operation. Plenty of warehouses have bought handhelds, scanned diligently for a year, and found the mis-ship rate roughly where it was — because scanning by itself is just a faster way to type.
What changes an operation is what the scan enables: work that is directed rather than remembered, moves that are verified rather than asserted, and an inventory record that is true mid-shift rather than only after the close. Each of those has a different cost and a different payback.
Past "we scan things now"
Paper picking runs on human memory and human care. A picker takes a list, walks the building in whatever order makes sense to them, identifies items by recognising them, writes quantities down, and hands the sheet to someone who keys it in later. Every step is a judgement call, and every judgement call has an error rate.
Those error rates are not distributed evenly. Your most experienced picker has one you would be happy with; your newest hire does not. The difference is knowledge that lives in one head, is written down nowhere, and walks out of the building at retirement. Directed, verified work moves that knowledge into the system — which makes the operational case a capacity and continuity case, not just an accuracy one.
| Step | Paper picking | Scan-driven picking |
|---|---|---|
| Receiving | Count, mark up the packing slip, key it later | Scan against the PO, discrepancy flagged at the dock |
| Putaway | Picker chooses a location that looks free | System directs a location; the location is scanned to confirm |
| Travel | Route chosen by the picker | Sequenced by location, so the route is walked once |
| Item identification | Visual recognition of the label | Scan must match the expected item |
| Quantity | Written, keyed later by someone else | Entered once, at the shelf, by the person who picked it |
| Packing | Spot check, if there is time | Second scan verifies every line against the order |
| Inventory update | At end of day, or next morning | At the moment of the transaction |
| Counting | Annual shutdown or a weekend | Cycle counts folded into normal shifts |
Receiving is the upstream root cause
Most picking errors are not created in picking. They are created at the dock, weeks earlier, and picking is simply where they become visible.
If a receipt is keyed as the wrong item, every later pick of that item finds the wrong thing in the location — and the picker, working from a system that says it should be there, believes the system. If a receipt lands in a location nobody recorded, the stock exists but is invisible, which produces the classic distribution absurdity of buying an item you already own. If a quantity is received in the wrong unit of measure, the error propagates into availability, into replenishment, and finally into a promise you cannot keep.
Scan-verified receiving fixes this at the point of entry. The receiver scans against the purchase order, the system compares what arrived to what was expected, and discrepancies are raised while the carrier is still at the dock. That is a far cheaper place to resolve a shortage than a customer phone call three weeks later — and the reason receiving should be the first process you convert, not the last.
An error caught at the dock costs a two-minute conversation. The same error caught at the customer costs a return, a re-pick, a credit, and a piece of your credibility.
Directed putaway and the pick path
Putaway that the system decides
Directed putaway means the system nominates the location for received stock and requires the operator to scan that location to confirm the drop. Two things follow. First, location accuracy stops depending on whether anyone wrote it down — every unit in the building has a recorded home, and the recording happened at the moment of the move. Second, the placement logic becomes a lever you can pull: fast-moving items near the pack stations, slow movers up high, related items grouped so a common order is picked in one pass.
Without directed putaway, placement drifts toward whatever was convenient on the day. The result is a warehouse whose layout reflects a decade of individual decisions rather than a demand pattern, and pickers who compensate with distance.
Sequenced picking versus wandering
The second half is the pick path. Paper picks are typically listed in order-entry sequence, which bears no relationship to the physical layout, so the picker either walks the list as written or re-sorts it in their head. Experienced staff re-sort; new staff walk it as written and cover the aisle four times.
Pick-path sequencing presents the lines in physical order, so the route is walked once, in one direction. The gain shows up as travel time, which in most distribution warehouses is the largest single component of a pick. It also shows up as consistency: when every picker walks the same route for the same order profile, throughput per labour hour becomes predictable enough to plan against, and a seasonal hire ramps in days rather than weeks.
Scan verification: the biggest single lever
If you change only one thing, change this. Scan verification means the item barcode must match the expected item before the line can be confirmed — at the pick, and again at pack.
It matters most because of the nature of the errors it eliminates. Look-alike items — the same part in two lengths, two voltages, two finishes, sequential part numbers on adjacent shelves — account for a disproportionate share of mis-ships, and they are exactly the errors visual checking is worst at catching. A person reading a shelf label in the fourth hour of a shift sees what they expect to see. A scanner does not.
Verification at pack adds an independent check on a different person's work, catching the class of error pick verification cannot: the right items in the wrong carton, or a line short-shipped because a case was left on the cart. Two verification points, by two people, is a real control. One is a partial one.
Cost a mis-ship properly before deciding how much verification is worth. It is not the price of the item: it is outbound freight, return freight, the credit, the re-pick, customer service labour, the second delivery, and the customers who quietly reduce their order frequency afterwards. Priced honestly, verification often justifies itself on freight alone.
Live decrements and cycle counting
When transactions post as they happen rather than in an overnight batch, the number on the sales floor stops being an educated guess. A rep quoting availability at 11am is quoting what is there at 11am, not what was there at close of business yesterday minus whatever has moved since. That removes a whole category of broken promise, and it is what makes the availability data in inventory management worth trusting while a customer is waiting.
Live posting also makes cycle counting practical — counting a small, prioritised slice of the building continuously during normal work instead of shutting down once a year. It only works when the system can tell an operator exactly what to count, accept the result at the shelf, and reconcile without a batch window. Handhelds make that a five-minute task between picks rather than a project.
The advantage over an annual count is not just the avoided downtime, real as a lost weekend is. An annual count tells you the size of your error once a year, after the damage. Cycle counting finds a discrepancy days after it happens, while the cause is still recoverable and the pattern still legible. Where the exceptions cluster — a location, a shift, an item class, a supplier — is the actual diagnostic, and that is the kind of question PrismView™ exists to answer.
Prerequisites, and the change nobody plans for
Four things that must exist first
- A clean location scheme. A consistent, human-readable convention — aisle, bay, level, position — applied across the whole building, including the odd corners: bulk storage, the mezzanine, returns, the will-call rack, the yard. Locations you do not name will hold stock you cannot find.
- Labelled bins. Printed, scannable, durable and reachable by an operator holding a device in one hand. Unglamorous work that takes longer than anyone estimates.
- Item barcodes that actually exist. Some items arrive with a scannable manufacturer barcode, some do not, and some carry a barcode encoding something other than the number you use. Decide and label for the second and third groups before go-live, not during it.
- A training period with real time in it. Not a lunchtime demonstration. Time on the floor, on the device, with support standing next to people while live orders go out.
The crew that has picked from paper for fifteen years
This part gets underestimated, and it has nothing to do with technology. Your most experienced pickers are your fastest today. On day one with a scanner they will be slower than the newest hire, in front of everyone. That is a hard thing to ask of someone whose standing rests on being quick, and handled carelessly it produces exactly the resistance you would expect.
Four things help. Say out loud that throughput will dip and that nobody is being measured on the dip. Involve the experienced crew in designing the location scheme and pick sequence — they know which routes work and will find the flaws in your layout logic faster than a consultant. Convert one process at a time, so each new habit lands on top of a stable one. And be explicit that verification is not surveillance: the scan exists to catch look-alike parts, not to build a case against a person. If the crew believes otherwise they will work around it, creatively.
Centerprism handles this through Wireless Warehouse Management: handheld scanning for picking, packing and receiving, every transaction posting to Microsoft Dynamics GP in real time, bin locations managed inside the ERP, pick sequences optimised for your layout, and cycle counts replacing the annual physical. Mobility extends the same live posting to the rest of the operation, so a warehouse update and a field lookup work from one record. The Wireless Warehouse datasheet (PDF) and the Mobility datasheet (PDF) cover both, and the rest of the library is in resources.
Readiness checklist
Work through this before you order hardware. Every "no" is a task, and tasks are cheaper before a go-live date exists than after.
- Every storage location in the building has a unique identifier, including bulk, mezzanine, returns and will-call.
- Those identifiers are physically labelled, scannable and reachable.
- Wireless coverage has been surveyed in the aisles and at floor level, not just from the office.
- Every active item either has a usable barcode or has an owner assigned to create one.
- Units of measure and conversion factors have been validated against how the item is actually stocked and sold.
- You have decided which processes convert first, and in what order.
- You know your current mis-ship rate and your current cycle-count variance, measured before the change, so you can tell whether it worked.
- Named supervisors are accountable for scan compliance on each shift.
- A realistic training window is in the plan, and the throughput dip is budgeted rather than denied.
- The go-live does not land in your peak season.
That last point is the one distributors most often override, and most often regret. If you want help working through the list against your own building — or a walkthrough of how a scan-driven pick actually behaves for wholesale distribution — request a demo or talk to our team about implementation.
