How to run a sawmill stocktake in New Zealand

If you run a sawmill or timber yard in New Zealand, you already know how stocktakes go. Someone announces the count is happening next Tuesday. Half the yard groans. The other half suddenly remembers they have something urgent to do in the workshop. Eight hours later you have a clipboard full of tallies that may or may not reflect reality, and the spreadsheet reconciliation takes another full day.

It doesn’t have to work this way. A well-planned stocktake at a medium-sized NZ mill should take a single shift for the count itself, with reconciliation wrapped up the following morning. The difference between an all-day slog and a clean count comes down to preparation, method, and knowing where the common traps are.

This guide covers the full process from start to finish.

Before the count: preparation that actually matters

The single biggest factor in stocktake accuracy is what you do in the 48 hours before anyone picks up a scanner or clipboard. Most stocktake errors trace back to poor prep, not poor counting.

Freeze movements

Stop all pack movements for the duration of the count. That means no dispatches, no forklift relocations, and no new packs coming off the saw or out of the kiln. If you run a yard where production can’t stop, designate a physical “quarantine zone” for any packs created or moved during the count and exclude them from the tally.

The worst stocktake errors happen when packs are being moved while someone else is counting. A pack gets counted at Station A, then a forklift driver shifts it to Station B, and the counter at Station B counts it again. You end up with phantom surplus that doesn’t exist.

Reconcile in-flight packs

Before the count, reconcile any packs that are in a transitional state. Packs marked as sold but not yet dispatched need to be clearly flagged. Packs sitting on a truck that hasn’t left yet need a decision: are they in your count or not? Packs that have been dispatched but not yet invoiced should already be marked as gone in your records.

The goal is to have a clean expected inventory list before you start counting. Every pack in your system should have a clear status: in stock, sold (awaiting dispatch), or gone.

Tidy the yard

This sounds basic, but it saves hours. Walk the yard the day before and make sure every pack is accessible and its barcode label is visible. Packs buried three deep against a fence with labels facing the wall are going to slow everything down. If you have packs without labels (it happens), flag them with a temporary marker so the counting team knows to record them as “unidentified” rather than skipping them.

Print or prepare your expected list

Whether you use software or a spreadsheet, generate a list of every pack you expect to find in the yard. Sort it by station or location. This becomes your checklist during the count and your reconciliation baseline afterwards.

If you’re working from a spreadsheet, include columns for: barcode/pack ID, species, grade, treatment class, dimensions, location, and a checkbox for “found”. See the pack label guide for a full breakdown of what each field should contain. If you’re using a scanning system, the stocktake feature handles the classification automatically.

Counting methods

There are three common approaches to the physical count. Which one suits your operation depends on yard size, team size, and whether you have scanning equipment.

Scanner-based counting

The fastest and most accurate method. Operators walk the yard with handheld PDA scanners in “count” mode, scanning every pack barcode they encounter. The scanner records each scan with a timestamp and the operator’s device ID.

The advantage here is automatic classification. The system compares each scanned barcode against the expected inventory and categorises it in real time: found (exists and expected), unexpected (scanned but not in current records), duplicate (already scanned by someone else), or unknown (barcode not in the system at all).

Scanner-based counting works well even in yards with patchy mobile coverage. A good offline-first system stores scans locally on the device and uploads them when connectivity returns. The operator doesn’t need to wait for a network connection.

Tag-and-tally

A middle-ground approach. Operators carry a clipboard and walk each row, checking each pack against a printed list. As they verify a pack, they tick it off and physically mark the pack with a chalk line or colour-coded tag to indicate it has been counted.

The chalk/tag step is important. Without a physical mark on the pack, you have no way to verify from the yard that a given pack was actually sighted versus just ticked off on paper.

Full manual count

The slowest method, but sometimes necessary for yards with no barcode infrastructure. Operators record every pack they encounter with its location, species, dimensions, and any identifying marks. This is then reconciled against whatever records exist.

If you’re doing a full manual count, work in pairs. One person reads the pack details, the other writes. This reduces transcription errors significantly.

During the count

Work in zones

Divide the yard into clearly defined zones and assign each zone to a specific counter or pair. Zones should follow natural boundaries: rows between stacks, station boundaries, shed vs outdoor, etc. Each counter is responsible for their zone only.

The zone approach prevents two people from counting the same pack and prevents gaps where neither person counted a section because each assumed the other would.

Handle treatment segregation carefully

One of the most common stocktake errors in NZ sawmills involves treatment-class mix-ups. A pack of H3.2 CCA-treated radiata looks almost identical to a pack of H4 CCA-treated radiata from a distance. The green tint is the same. The timber dimensions might be the same. The only reliable identifier is the label.

If labels are missing or damaged, record the pack as “treatment unverified” rather than guessing. Guessing at treatment class during a stocktake creates false precision that can cause real problems downstream, especially if you’re reporting stock to customers who are ordering specific treatment classes.

Record quantities consistently

Decide upfront whether you’re counting in piece count, lineal metres, or cubic metres. Stick with one unit for the entire count. You can convert afterwards, but mixing units during the count guarantees reconciliation headaches.

Most NZ sawmills count in pieces during the physical walk and convert to cubic metres for reporting. A standard conversion for common radiata dimensions:

A pack of 100 pieces at 90x45mm x 4.8m = 100 x 0.09 x 0.045 x 4.8 = 1.944 m3.

Whatever your approach, make sure the conversion factors are documented and consistent. Rounding errors compound quickly across hundreds of packs.

After the count: reconciliation

This is where the value of a stocktake actually materialises. The count itself is just data collection. Reconciliation is where you find the problems.

Classify every discrepancy

Compare your physical count against your expected inventory. Every pack should fall into one of four categories:

Found: Pack exists in your records and was physically counted. Good.

Missing: Pack exists in your records but was not found in the yard. This is the one that costs money. Missing packs could mean theft, dispatch errors (pack left on a truck and wasn’t recorded), misplacement (it’s somewhere in the yard but wasn’t found), or a data entry error when the pack was originally created.

Unexpected: Pack was found in the yard but doesn’t exist in your records. This usually means a pack was created but never entered into the system, or a returned pack wasn’t re-entered after coming back from a customer.

Unknown: A barcode was scanned or a pack was found that doesn’t match anything in any system. Could be an old pack from before your current tracking system, a pack from another site, or a labelling error.

Investigate missing packs before adjusting

The temptation after a stocktake is to adjust your records immediately to match the physical count. Resist this. Missing packs should be investigated before you write them off.

Check the dispatch records for the past 30 days. Was the pack dispatched but not marked as gone? Check with forklift operators. Was the pack moved to a station that wasn’t counted? Check with the treatment plant. Is the pack currently in treatment and temporarily off-site?

Only after investigation should you adjust records. A stocktake that writes off 50 “missing” packs without investigation is doing half the job.

The variance report

Once reconciliation is complete, generate a summary that captures: total packs expected vs total found, variance by species and treatment class, variance by location/station, and a list of every missing, unexpected, and unknown pack with investigation status.

This report is your stocktake deliverable. It should be stored for reference and compared against future stocktakes to identify patterns. If you consistently find missing H3.2 packs at Station C, that’s a process problem worth digging into.

GST and IRD timing

For NZ businesses, the timing of your stocktake matters for tax purposes. Timber inventory is trading stock under the Income Tax Act 2007, and its value at balance date affects your taxable income. If your balance date is 31 March, running your stocktake in the last week of March gives you the most accurate figure for your tax return.

If you can’t time the stocktake to balance date, run the count as close to it as practical and document any known movements between the count date and balance date. Your accountant will thank you.

How often to count

The answer depends on your operation size and how much stock movement you have.

Small mills (under 5,000 m3/year): Twice a year is usually sufficient, with one count aligned to your balance date.

Medium mills (5,000-20,000 m3/year): Quarterly is ideal. Monthly is better if you have the capacity.

Large mills (20,000+ m3/year): Monthly counts of the full yard, with weekly spot-checks of high-turnover stations.

If you’re using a real-time inventory system with barcode scanning, the frequency matters less because your records stay closer to reality between counts. The stocktake becomes a verification step rather than the primary method of knowing what you have. That said, even with the best system, physical counts catch things that digital records miss: damaged packs, labels that have fallen off, packs that were moved by someone who forgot to scan.

Common mistakes

Counting during production. Already covered above, but it’s worth repeating. A stocktake during active production is a stocktake with errors.

Not counting what’s in the kiln or treatment plant. Packs in the kiln or at the treatment plant are still your inventory. Include them in the count or document them as a known exclusion.

Ignoring broken or damaged packs. A pack that’s been damaged and is sitting in a corner waiting for a decision is still a pack. Count it, flag it, and deal with the write-off as a separate process.

Rounding to save time. “That stack looks like about 40 packs” is not a count. Estimate-based stocktakes consistently undercount by 5-15%, which compounds into real money over a year.

Not comparing to previous counts. A single stocktake tells you what you have today. A series of stocktakes tells you where your process is leaking. Track your variance percentage over time. If it’s trending up, something in your operation is getting worse, not better.