Scanner flat, offline or data lost on count day: a backup plan

Sooner or later a device fails during a count. The battery dies at three in the afternoon, the Wi-Fi in a market hall drops out, or a handheld restarts and the last hour of scans cannot be found. None of this has to be a disaster. What decides how bad it gets is the size of the piece you can lose: if everything lives on one device until the end of the day, a failure costs the day. If you save after every zone, it costs one zone, and with a paper log of finished bins it can shrink to a single bin.

Before the day: prepare devices and people

• Charge every device the night before and start with a full battery. For long sessions bring spare batteries or power banks.
• Take more devices than teams, roughly one spare for every five in use.
• Number the devices (D1, D2, D3) and note on a sheet who holds which one.
• Check free storage and make sure the date and time are right on every device. Scans with a wrong clock cannot be put in order later.
• Test the whole chain on a few bins at the real location: scan, save, export, open the file on a computer. An office test says little about a hall with weak signal.
• Assume there is no network. Work in a way that stores scans on the device and exports them afterwards.

During the count: save in small pieces

Agree one rule for everybody: export after every zone or every 60 minutes, whichever comes first. Name the file with the device, zone and time, for example D3_A_1540. Send each file to the coordinator through a chat group or by email, and keep the copy on the device until the count has been accepted. At the same time the zone leader ticks off finished zones on a paper sheet, so everyone knows which parts are safe.

When something goes wrong

• Battery dies or the device freezes: swap in a spare, write it on the paper sheet and rescan the bin that was in progress. Everything exported earlier is safe.
• No connection: carry on offline and export when the zone is finished. Do not stop work to chase a signal.
• Records lost or a file will not open: recount only the zones that were not yet exported, using the paper sheet to see which they are.
• Duplicates after a restore: the same bin appears in two files. Keep the later, complete one and remove the other during cleaning. Do not add them together.
• Several devices fail: switch to paper sheets printed beforehand, one per bin, with columns for SKU, quantity and initials. Two people key them in later, separately, and the two results are compared.

An example with numbers

Suppose four teams each export every hour. At 15:40 one device freezes and cannot be recovered, and its last export was at 15:00. The team loses 40 minutes of work in a single zone, perhaps three or four bins, and recounts them with a spare device in the next hour. Without hourly exports the same failure would have cost that team the whole day, perhaps 40 bins, and the count would have run into the night. The figures are only an illustration, but the proportion is the point.

Afterwards: keep every file

Keep every exported file, including partial ones, until the adjustments have been approved. A file that looks incomplete is sometimes the only evidence for a particular bin. Note in the count log which devices failed, when and in which zone.

Paper is still a tool

A paper sheet feels old-fashioned next to a scanner, but it is the only part of the plan that does not depend on a battery. Print a small stack per zone, with the bin codes already on them, and put them in a folder at the zone entrance. For valuable zones, ask the counter to write the bin code and its total on the sheet as each bin is finished, which takes a few seconds. Then even a complete device failure costs at most the bin in progress, because the paper holds the total of every finished bin. If the sheets are never needed, they cost a little paper. If they are, they save the afternoon.

InventoryHub's handheld scans record the bin, time and operator with every item, and bins that do not tally are recounted straight away, so a gap in the records shows up while the team is still on site.