Handheld terminal stock counting means scanning each product's barcode in the warehouse or store with a portable reader, entering the quantity on the device and sending the result to your system without paper or re-keying. Set up properly, it shortens the count, removes the transcription errors that come from re-keying paper sheets and produces a variance report the moment counting ends.
Why paper counts always stall in the same places
In a traditional stock take the team spreads out across the racking with printed sheets. Quantities are written down by hand and, that evening, someone types them into Excel or the ERP. Errors creep in at every step: a figure on the wrong line, illegible handwriting, a bay counted twice, a corner never counted at all.
When the count is over the variances are visible, but their causes can no longer be traced. Nobody knows who counted which bay, and if goods were received or shipped during the count the figures have already moved. So many businesses count once a year because they have to, and do not trust stock figures in between. Most of the issues in our article on outgrowing spreadsheets turn up in count sheets too.
Under Türkiye's Tax Procedure Law, businesses keeping balance-sheet-basis books take an inventory at year end, and the result feeds directly into the financial statements. A reliable count is a need for the finance team as much as for operations.
What inaccurate counts cost
Count variances look small but they add up. Data from US retail shows the scale of the problem, though the figures should not be applied directly to Türkiye:
According to the National Retail Federation's National Retail Security Survey 2023, the average US retail shrink rate rose to 1.6% in FY2022, representing $112.1 billion in losses.
Theft is not the only cause of stock discrepancies; record errors from manually keyed data play their part too. A 2018 study by the Auburn University RFID Lab and GS1 US found that when automatic identification (RFID) was not used, 69% of orders shipped from brands to retailers contained data errors. Every figure written on a count sheet and keyed in later is exposed to the same kind of mistake.
Goods-in is moving to electronic documents quickly as well. According to the Turkish Revenue Administration's 2025 Annual Report, the number of taxpayers using e-Dispatch Notes (e-İrsaliye) rose from 402,985 in 2024 to 632,288 at the end of 2025. Receiving goods that arrive with an electronic dispatch note on a handheld terminal helps catch many errors at goods-in, before they turn into count variances; our guide to e-invoice integration in Türkiye explains how these documents flow into your systems.
What to look for in a handheld terminal
A handheld terminal is a rugged mobile computer with a built-in barcode scanner. Most current models run Android, and the counting app runs directly on the device.
| Criterion | Why it matters |
|---|---|
| Scanner type (1D/2D) | QR and DataMatrix codes need a 2D imager |
| Scan range | Long-range scanners speed up work on high racking |
| Ruggedness and IP rating | Resistance to drops, dust and moisture; operating temperature for cold stores |
| Battery | Must last a full shift, with easy battery swaps |
| Connectivity | Where Wi-Fi coverage is patchy, the device should work offline and sync later |
| Screen and keypad | Physical keys make a real difference when staff wear gloves |
We explain what dust and water resistance actually mean in our guide to IP65 and IP67 ratings. Where you need bulk reads or line of sight is awkward, RFID may be a better fit than barcodes; we compare the two in RFID inventory and asset tracking.
Handheld terminal stock counting: an eight-step framework
- Choose the type of count. A full count covers the whole warehouse at once; cycle counting splits it into sections and counts one each day or week. Counting high-value or fast-moving lines more often (an ABC approach) puts the effort where it matters.
- Label every location. Each bin needs a readable location barcode. Scan the location first, then the product, so the count also corrects where each item is.
- Clean up product barcodes. Before counting, fix items without barcodes, two products sharing one code, and mixed-up case and unit barcodes. Units of measure (case, pack, each) must be defined in the system.
- Freeze or separate movements. In a full count, receiving and dispatch stop in the area being counted. In a cycle count, movements pending at the time of counting are flagged separately.
- Count blind. The counter should not see the system quantity on screen. Anyone who sees the expected figure unconsciously counts towards it.
- Recount variances with a second team. Variances above a threshold go to a different person for a recount. Only if the second count confirms the difference is the record corrected.
- Record the cause of each variance. Receiving errors, misplaced stock, waste, loss or data entry mistakes should be reported separately so the process can be fixed.
- Post results to the ERP under control. Adjustments reach the accounts only after approval. A counter should never be able to change stock quantities without sign-off.
Why cycle counting works better
A big annual count tries to find a whole year's errors in a single day. With cycle counting, a variance is caught soon after it appears, while its cause can still be traced. The warehouse keeps running and the count team works through a few locations a day.
Cycle counting is most effective alongside a warehouse management system (WMS) that tracks locations and batches. For products with expiry dates, scanning batch and date during the count also shows whether FIFO and FEFO inventory methods are really being followed on the floor.
Common counting mistakes
A handheld terminal removes the transcription errors that paper causes, but it does not fix weak processes by itself. These are the mistakes we see most often on site:
- Confusing cases with units. Scan a case barcode and key in units, or the other way round, and the count is out by a factor of ten or twenty. The app must know the unit conversions and show clearly whether the scanned code is a case or a single item.
- Sending two people to the same bay. If tasks are not assigned by location, one bay gets counted twice and another not at all. Each location should be assigned to one person and locked once complete.
- Forgetting movements during the count. If stock ships from an area being counted, the variance is not real. When counting starts, record which movements are on hold.
- Closing variances without a reason. Stock corrected in a single "count variance" line will produce the same variance at the next count.
- Skipping training. Staff picking up a device for the first time on count day work slowly and make mistakes. A short practice count beforehand noticeably improves the real one.
Most of these mistakes can be blocked on the device if the app is designed well. A screen that refuses a product scan until a location has been scanned, warns on unusually high quantities and locks completed locations takes a lot of work off the count supervisor. That is why the counting rules should be written with the people who know the warehouse best.
Using count data well: where do variances come from?
Count results are not just for correcting stock; they show where the process breaks. Which product group, which location and which shift produce variances? Master data problems, such as the same item opened under two codes, are a frequent cause; our guide to data quality and duplicate records explains how to clean them up.
In warehouses where pickers need their hands free, voice picking is worth considering alongside handheld terminals. In store chains, accurate counts are the foundation of retail inventory optimisation.
Accurate stock records also shape the till experience: an item missing from the system cannot be scanned at a self-checkout kiosk and leaves the customer waiting for help. That is why counting discipline needs to be in place before a store moves to self-service.
Barcode counts with handheld terminals are also the backbone of medical equipment inventories in hospitals, with BLE tags or UHF and HF RFID tags added to devices that move between departments. We cover that mixed set-up in our guide to hospital medical equipment tracking.
In retail chains, count results feed transfer and ordering decisions directly; we cover that link in our guide to retail chain store management systems.
How we deliver counting projects at Digital Bridge
We treat a counting project as process and software together, not a device sale:
- Discovery and requirements analysis. We review your warehouse or store layout, product and barcode structure, current counting method and ERP. Scope, phases and cost are confirmed in a written proposal.
- Device selection. We choose handheld and industrial data terminals to suit the environment; cold stores, outdoor yards and shop floors may each need a different model.
- App development. We build counting, receiving, transfer and dispatch screens around your workflow as a mobile app. Blind counts, recounts and variance approval are built into the app.
- Integration. We post count results to your ERP through API integration, once variances have been approved.
- Warehouse management. Where location, batch and expiry tracking are needed, we extend the counting app into a warehouse and inventory management system (WMS).
- Pilot and roll-out. We start with cycle counting in one warehouse or zone and extend to other sites based on the results. We work remotely and on site across every province in Türkiye.
For more on devices built for the field, our IoT device development process guide is a good starting point.
Next step
List the ten products with the largest variances in your last count and where they are stored. Then get in touch through our contact page; we will review your counting process with you and show, with a concrete plan, how cycle counting with handheld terminals would work in your warehouse. For more on field hardware, visit our IoT & Hardware hub.