**Run a stock audit by narrowing it to one site, walking the floor with a phone,
and scanning the label on each item. Nothing is called missing until you sign the
audit off**, so you can walk in whatever order suits the building.
Most companies audit assets once, find it took a weekend, and never do it again.
That is a labelling problem rather than an auditing problem.
Why the clipboard version is so slow
Counting without labels means reading a serial number off the underside of a
laptop, finding it in a list of four hundred rows, and ticking it. Roughly ninety
seconds per item, most of it spent searching rather than looking.
Four hundred assets at ninety seconds is ten hours. That is the weekend, and it
is why the second audit never happens.
Scanning a label is about three seconds, and the searching is done by the
database rather than by a person squinting at a spreadsheet on a phone. The same
four hundred assets take under an hour, and most of that is walking between
rooms.
Print the labels first
This is the step that makes everything else possible, and it only has to happen
once.
Wurxa prints barcode or QR labels for the whole register, laid out on an A4
sheet. Both encode the asset tag, so a scanner gun and a phone camera resolve to
the same asset. Stick them somewhere that will still be visible when the item is
in use: the underside of a laptop is findable, the back of a monitor against a
wall is not.
For vehicles and machinery, put the label where somebody standing next to it can
read it without moving anything. The audit is only fast if the label is.
Narrowing the audit, which is the part people get wrong
Create the audit for one site, and in a large building, one location within
it.
This matters more than it sounds. An audit builds its expected list from the
register, and anything on that list you never scan is recorded as missing when
you sign. An audit of every site at once cannot be walked in a single pass, so
signing it marks a building's worth of perfectly present equipment as missing.
Narrowing is what makes an audit finishable, and a finished audit is the only
kind that tells you anything.
Walking it
Open the audit, start scanning. On a phone the camera reads the label; if the
camera is awkward for a particular item, type the tag instead and press enter.
Three things make the walk work:
The scan box keeps focus and clears itself. You are holding a phone in one
hand and an object in the other. A form that needs a tap between scans is a form
that gets abandoned.
Each result has its own sound. A rising tone means recorded, a flat repeat
means look at this one, a low buzz means it did not register. You are looking at
the shelf, not the screen, so the tone is the feedback and the text is what you
check afterwards.
Nothing is called missing during the walk. An unscanned line stays "not seen"
until you sign. Scan the far corner first if you like; the order does not matter.
What the audit tells you at the end
Three categories, and the middle one is the valuable one.
Moved and unexpected are the findings. Equipment migrates between branches
for entirely sensible reasons and the register is the last thing anybody updates.
An audit that only told you what was missing would throw away the more common
discovery, which is that things are present and in the wrong place.
Signing it off
Signing freezes the audit. Missing items are recorded against the register at
that moment, and nothing about it can be changed afterwards, including further
scans.
That finality is deliberate. An audit you can edit later is not evidence of
anything, and the reason to run one is usually so that somebody can rely on the
result.
Sign it when you have finished walking, not when you have finished investigating.
The four items you cannot find are a separate piece of work, and holding the
audit open while you chase them means the count reflects a week of movement
rather than a moment.
How often
Quarterly per site is enough for most companies. Monthly for high-value or
high-turnover stock. Annually is common and slightly too rare, because a year of
undocumented movement is hard to reconstruct.
The honest test is whether the last audit changed anything. If it found nothing,
audit less often. If it found a dozen moved items, the register is drifting
faster than you thought and the interval is too long.
What to do with what the audit found
An audit that produces a list nobody acts on is a slower version of not auditing.
Three actions cover almost everything it turns up.
Moved items: update the register, do not investigate. Something is at branch
B and the register said branch A. That is not a mystery, it is a move nobody
recorded. Change the location and move on. Investigating every relocation is how
audits acquire a reputation for being punitive, after which people stop
cooperating with them.
Unexpected items: add them. A tag scanned that the expected list did not
contain is usually equipment that arrived from elsewhere. It is now recorded,
which is a better position than before you walked.
Not seen: ask once, then decide. Give it a week. Most turn up in a drawer, in
somebody's bag, or at the repair shop. What remains after a week is genuinely
gone, and marking it lost is more useful than leaving it on the register looking
present.
The last one matters for the register's credibility. A list that contains things
nobody has seen for two years teaches people that the list is fiction, and once
they believe that, they stop scanning.
Who should actually do it
Not the person who maintains the register.
Somebody auditing their own records is checking their own work, and they will
unconsciously walk the route that matches what they expect to find. Give it to
somebody who works in that building and does not own the system. They will scan
what is in front of them rather than what they are looking for.
It also takes an hour, which is a reasonable thing to ask of somebody who is not
otherwise involved with asset management at all.
The four answers, and why two of them get confused
Building the audit engine forced a distinction we had not expected to matter, and
it turns out to be the difference between a stock take people act on and one they
argue about.
An audit asks one question per asset: was it where we said it was? There are four
answers, and they are not symmetrical.
Found. Expected here, seen here. The boring majority, and the point of the
exercise.
Missing. Expected here, never scanned. The one everybody thinks an audit is
for.
Moved. Expected somewhere else, but seen here. The asset is yours, it is
fine, and your register is out of date.
Unexpected. Scanned here, and there is no record of it at all. The tag may
not be yours.
Moved and unexpected look alike on a phone screen and they are completely
different problems. A moved asset is a data fix: update the location and carry
on. An unexpected tag is a question. Did somebody buy this outside the process?
Is it a supplier's equipment sitting in your store room? Is it an item that was
written off and never actually disposed of?
Systems that collapse the two into "discrepancy" produce a number nobody trusts,
because the pile contains both trivia and genuine surprises and there is no way
to tell them apart without walking the floor again.
One more thing the code had to get right: **absence is a property of the whole
audit, not of any one scan.** You cannot mark an asset missing while the count is
still running, because the person doing it may simply not have reached that shelf
yet. Missing is only knowable at the end, when the scope is closed.
That sounds obvious written down, and it is the single most common way a
half-finished stock take generates a panic about equipment that was two metres
away. The practical version, whatever tool you use: scope it, finish it, then
look at the exceptions. An audit read halfway through is worse than no audit,
because it produces confident wrong answers.
Put the workflow into practice.
Start with every Wurxa feature for 14 days. No credit card required.
