Case studies — AU Logistics

Dangerous goods · Australia to New Zealand · 2024Two countries, one pallet, and a word nobody had defined

AU Logistics — case study 01 — Warehousing and distribution

The freight was never the problem.

An Australian fragrance house had finished product, a distributor conversation in Auckland and no route. The product is 90% ethanol. That makes it dangerous goods Class 3, and that one fact decides who may store it, who may fly it, what paperwork travels with it, and what it costs to put one bottle in one buyer’s hands.

13×Between the two routes to marketLanded cost per unit, distributor versus direct to the buyer.

70×Hidden in one undefined wordThe swing between the two readings of “per unit” on a single line.

5,040Units the chain was priced againstTheir pallet, their cartons, their unit weight — not a generic one.

2 → 1Vendors in the destination countryOne handover removed before a single carton shipped.

Two routes to market, one pallet

Landed cost per unit, indexed

Same product, same pallet, same aircraft. Only the buyer changes.

Sold to a distributorpallet in, pallet out1×

Sold to the buyerpicked and couriered, one at a time13×

Airfreight

Dangerous-goods documents and handling

Pick and courier in New Zealand

Indexed to the bulk route. Built from the schedules the Australian and New Zealand parties actually issued, applied to the client’s own pallet. A rural delivery pushes the one-by-one figure to roughly 27×.

The line that decided it

“Dangerous goods handling — per unit”

One line. Two readings. Nobody had said which.

If a unit is a carton

cartons handled. A rounding error on the shipment.

If a unit is a bottle

5,040

items handled. Near enough the airfreight bill again.

A 70-fold swing sitting inside one undefined word, on a line neither party had queried. The same ambiguity sat in the New Zealand pick fee. You do not find this comparing headline rates. Both quotes read the same at the top.

The rate card cannot tell you this. Your shipping calendar can.

What we did

Asked one question before quoting anything. Bulk to a distributor, or one at a time to the buyer? Licensed dangerous-goods stores are built for pallets. They will pick singles. It is a different operation and it prices like one.
Priced their pallet, not a generic one. Their dimensions, their carton count, their unit weight, run line by line through every schedule in the chain.
Built the chain from licensed parties in both countries — Australian dangerous-goods storage and handling, airfreight carrying Class 3 surcharges, a New Zealand hazardous-substances store, border and biosecurity processing, and last-mile courier including the rural band. The dangerous-goods declarations are the consignor’s and are signed by trained, certified people; our job was to make sure the chain contained someone entitled to sign them.
Cut a handover nobody needed. The chain first ran through two New Zealand vendors. The forwarder volunteered that its own leg added complexity the job did not require, and the store confirmed it could run the country end to end.
Read the schedules for the words that decide the deal, not the rates at the top. An undefined “unit”. A per-kilo rate drawn from a weight bracket this shipment does not reach.

What you hold at the end

Landed cost per unit for both routes to market, built from real quoted rates against your real pallet, with the arithmetic shown so you can check it.
The frequency answer. Minimum charges in this chain do not care how often you ship. At one pallet a month they are a tax. At a pallet a fortnight they disappear into the volume.
A named, licensed party for every leg in both countries, with what each one is and is not licensed to do written down.
The questions to put back to the vendors in writing before you commit — the ones whose answers move the number by an order of magnitude.

The finding

One 155-gram bottle crosses the Tasman — storage, dangerous-goods paperwork, the flight, the border, a licensed store — for less than a tenth of what it then costs to pick that same bottle off a shelf and courier it across New Zealand. The international leg everyone negotiates turns out to be the cheap part. The rest of the landed cost was settled in a sales meeting, by choosing who to sell to, months before anyone asked a forwarder for a rate.

The detail

The consignment as specified

Perfume, 90% ethanol, transport class DG Class 3.
One pallet, 1150 × 1170 mm footprint, 1100 mm high, 785 kg gross.
72 cartons in 8 layers of 9. 70 units per carton at 155 g. 5,040 units on the pallet. One SKU.

The chain as assembled

Australia: dangerous-goods storage and handling, document preparation, and the transport documents required to tender Class 3 cargo to an airline.
Air: Class 3 acceptance with the surcharges that attach to it.
New Zealand: border and biosecurity processing, road movement into a licensed hazardous-substances store, storage under the applicable storage class, then either bulk release or pick and courier.

What we would ask the vendors before committing

Define “unit” on every per-unit line, in writing, on both sides of the Tasman.
Confirm the weight bracket the per-kilo rate is drawn from, and whether it holds at the actual chargeable weight of this shipment.
Confirm the storage minimum and the period it is charged over, against your intended shipping frequency rather than a full pallet position.
Confirm which entity holds which licence, and what happens to the cargo if the store cannot accept it on arrival.

Want the long version? The full case study carries the consignment detail, the chain as assembled and the questions we put to each vendor.

Pallet and classification are the client’s own, from their safety data sheet. The ratios are arithmetic on the rate schedules the Australian and New Zealand parties issued in October 2024. The brand did not proceed.

Air parcels · licensed depot · Brisbane gateway · 2024–2026The service standard the contract term could not buy

AU Logistics — case study 02 — Specifications and tenders

The site could hit the volume or hit the KPI. Never both.

A global marketplace was flying parcels into Brisbane on scheduled bellyhold and dedicated overnight charters, feeding the national postal operator’s international arm. Our team ran the licensed depot they landed in: recover from the airport terminals at first light, break down the unit load devices, scan, cage, segregate anything the agencies flagged, and have the building empty by close. The contract asked for every parcel scanned. That is where it got interesting.

~35kThe honest scanning ceilingPer day, with overtime and a bigger labour pool. The number nobody wanted to say.

40k+What the site was sized forThe design point the roster and the cost base were built around.

50–55kDocumented single-day peaksIncluding the largest Saturday on record for the site.

7 vs 18–24Months of term against months of paybackMeeting the standard at peak needed machinery the contract was too short to amortise.

The service standard, against the site’s own physics

Parcels per day

The contract asked for every parcel to be scanned. These three numbers are why that could not hold at peak.

Ceiling on 100% scanningwith overtime and an enlarged labour pool~35k

What the site was sized forthe design point the roster was built to hold40k+

Documented single-day peaksthe days the network actually delivered50–55k

Scanning every parcel and moving peak volume were different operations with different ceilings. Closing the gap meant mechanising. Mechanising needed 18 to 24 months to pay back, against a contract with seven months left to run. The service standard was not expensive — it was unbuyable on that term.

What actually bound the site, in the order it bound it

FIRSTPeople

Scanners and forklift drivers. Solvable: hire, roster, bring in labour hire.

THENFloor

Cubic space for unit load devices and cages. Not solvable by rostering. Procured in months, not shifts.

THENOutbound trucks

Full cages could not be evacuated and empty ones could not be supplied. A customer-provided input. The site stopped with staff on the floor and work waiting.

None of the three was the number in the contract. The last one was not even ours to fix.

A worked example of the same problem, from the same account

“Are you able to provide reverse logistics?”

One line, inside a four-part request covering footprint, systems, performance and customer service. It is the only mention returns get.

What we asked back, the same day

Do the returns arrive through the postal network, or do we collect them from the consumer?
Is it the whole catalogue, or only the items that came through us?

What those two answers decide

Whether this is a receiving dock or a national collection network.
Whether you hold two thousand line items or two hundred thousand.
Whether low-value goods travel back at all, or are received and written off where they land.
Between them, an order of magnitude on the price.

Neither question was answered, and no returns price was issued. That was the right outcome. The same request put the forward volume at three thousand orders a day rising to ten thousand, on a throughput-only rate with no storage component. At the dispatch rate the team had costed, that is somewhere between twenty-five and a hundred and twenty-five people on the floor — a range set entirely by questions the request never asked. A number quoted into that is not a quote. It is a liability with a decimal point.

We had the empty cages. We had the people. The site still stopped at ten in the morning.

What we did

Named the constraint each time it moved, instead of restating the headline volume. Over two years it was people, then floor, then outbound trucks — and the answer was different every time.
Costed the two processing modes separately. Bulk decant and parcel-level scanning are different operations with different ceilings. Averaging them hides the only number that matters at peak.
Tested the service standard against the contract’s own term. Holding 100% scanning through peak required mechanisation; mechanisation required 18 to 24 months to pay back; the contract had seven months left. That is an arithmetic answer, not an opinion.
Made the dwell clock symmetrical. Storage accrued against the customer when its own equipment sat uncollected, and against us when our truck missed a recovery window. Same accessorial, both directions, in writing.
Gave up our own margin where it bought nothing. On the second-city operation we held a head contract over a subcontractor. We collapsed it and introduced them to the customer directly rather than carry a margin on a rate rise.
Refused to price an undefined scope. Asked whether we could handle returns, we came back with the two questions that set the price rather than a number that would have had to move later.

What you hold at the end

A specification written in the units the site is bound by — throughput at each processing mode, floor area, dock and cage evacuation rate — not just parcels a day.
The capital question answered before you sign. Which service levels need a machine, what that machine needs to amortise, and whether your term can carry it.
Volume language that cannot be read two ways. Every figure with a unit of time attached to it, and a stated consequence when actuals diverge.
A dwell-clock schedule that runs both ways, so the party whose truck failed is the party who pays for the delay.

The finding

A service level is a capital decision wearing an operations costume. Scanning every parcel through peak needed machinery. The machinery needed about two years to pay for itself. The contract had seven months left to run. Past that point the standard cannot be bought at any price — not by us, and not by anyone bidding against us. So the number to check before signing is not the volume. It is whether your own term outlasts the payback on whatever your service standard quietly requires.

The detail

The daily cycle

Recover consignments from the airport cargo terminal operators on a runsheet sequenced by terminal, earliest-access first, because same-day recovery had a hard cut-off after which the shift change made the freight unavailable.
Break down unit load devices; either bulk decant or scan at parcel level; build into roll cages for the postal network.
Segregate anything held by the border or biosecurity agencies into a monitored, segregated area, and present it when the agency attended.
Release built cages to the customer’s collection vehicles. Dwell was measured in hours. The building was designed to empty itself daily.

Why the two processing modes mattered

Parcel-level scanning gives track-and-trace integrity and is what the service standard required.
Bulk decant moves peak volume and is what the floor allowed.
The gap between them is not a performance problem. It is a capital problem, and it has to be answered at the specification.

What we would put in the specification next time

Throughput stated separately for each processing mode, with the ceiling for each.
Floor area, dock count and cage evacuation rate as contract parameters, not operational detail.
A unit of time attached to every volume figure, and a stated remedy when actual volume diverges from guidance.
Any service level that requires capital, priced against the contract term, with who funds it and what happens to it at expiry.
Accessorials that run in both directions on the same clock.

Want the long version? The full case study carries the consignment detail, the chain as assembled and the questions we put to each vendor.

An air parcel contract run from a licensed Brisbane depot, 2024 to 2026. Volumes are the site’s own processing records. The depot licence, bond integrity and presentation of goods for inspection are depot functions; declarations and clearance determinations belong to the licensed broker and the agencies.

Project freight · four load ports · North Queensland · 2025–2026The freight was fine. The container was the problem.

AU Logistics — case study 03 — Project freight and network design

Nobody was going to fix this with a rate.

A modular housing manufacturer was importing prefabricated dwellings and building components out of four Chinese ports for a multi-thousand-home development in North Queensland. The supplier shipped on its own terms into a southern port. From there the boxes had about 1,100 km to travel north, an estate that could receive four containers a day, and then 1,100 km back to an empty park before the free-time clock ran out.

10–13Days under the carrier’s clockDischarge to de-hire on the forced cycle, including the queue at a four-a-day gate.

10–17Days of free time allowedThe spread across carriers on this lane, before a single kilometre was driven.

~$450Per container, worst windowAt a ten-day window on published tariffs. Nil at seventeen. Same cargo, same week.

4⁄5Of the saving came from the free decisionMoving the discharge port, not buying the containers.

Three discharge ports, one destination

Where the boxes could land, and what each choice cost in distance

The supplier had chosen the southern port. The site was 1,100 km away from it and about 200 km from two others.

Discharge port specified and priced

Destination estate

Road on-carriage

Long-haul backup lane, road or rail

Ocean legs from four load ports

The northern direct call was executed. The second northern port was priced and held as the schedule alternative but only one carrier called there. The southern lane stayed as the contractual backup and ran roughly 35–45% above the direct northern call, all in, per container, on the route matrix issued to the client.

The container cycle, before and after

The cycle the supplier’s shipping term forced Does not close

Discharge at the southern portsupplier’s chosen term Road north~1,100 km Unpack at the estatesite takes 4 boxes a day Dead-haul south~1,100 km, empty Return before free time10–17 days total

Two 1,100 km legs and a site-paced unpack, inside a window that varied from ten days to seventeen depending only on which carrier was booked.

The cycle we specifiedNo clock at all

Discharge at a northern port~200 km from site Sideloader to hardstanddirect, one box per truck Unpack at the client’s paceno window Box stays on the estaterepurposed on site No return legnothing to return

Shipper-owned boxes, built to seaworthy certification and reinforced so they could be reused on arrival. The container stops being a leased asset on a clock and becomes project material. Detention and de-hire are not negotiated down. They stop existing.

What the clock actually cost

Days the box was held, against the free time each carrier allowed

Discharge to de-hire on the forced cycle: 3–4 days from discharge to gate-out, two days on the road out, a day to unpack, two days dead-hauling the empty back, a day for the de-hire slot — plus nought to five days queueing at a gate that took four boxes a day.

Same cargo, same site, same week. Whether it cost nothing or several hundred dollars a box came down to which service happened to be booked. Costed on published Australian tariffs — about $180 a container a day for the first week past free time, rising to $300–$397 a day beyond two weeks (Hapag-Lloyd Australia effective 1 October 2025; ONE Australia valid from 1 April 2026; ex-GST). These are published tariff rates. An account shipping forty-odd boxes a month negotiates its own free time and would pay less.

Where the saving actually came from

Annual detention exposure, at the intended run rate

Three versions of the same programme. We modelled them back afterwards to see which decision had done the work.

Southern port, carrier’s boxthe cycle the supplier’s shipping term forced~$103k

Northern port, carrier’s boxsame box, same carrier, 900 km closer to site~$19k

Northern port, own boxthe container never goes backnil

Indexed to the forced cycle, equal-weighted across the four free-time windows the carriers quoted. Moving the discharge port takes out roughly four fifths of the exposure on its own, and costs nothing — it is a line in the purchase order, not a capital item. Buying the containers removes the remainder. Conditional on the intended 40–50 boxes a month being reached on the unchanged cycle with no change of carrier; seven containers have shipped to date, all shipper-owned, and no detention was invoiced on this programme.

The expensive fix was not the one that did the work.

What we did

Re-opened the port decision. The supplier’s shipping term had already chosen the discharge point, and it put every box 1,100 km from site. We specified and priced three ports all the way to the estate, and moved the programme onto a direct northern call about 200 km out.
Normalised four origins onto one cost stack. Ocean rates came from the carriers and origin agents; the work was putting four unrelated quotations into one comparable landed position per container.
Treated free time as a selection criterion. Ten days against seventeen is nearly a factor of two on exposure, decided purely by which service was booked, and it usually appears nowhere in the rate comparison.
Changed what the box is for. Shipper-owned containers, certified seaworthy and reinforced so they could be repurposed on arrival, delivered by sideloader straight to hardstand and retained on the estate. No empty repositioning, no de-hire event, no free-time clock.
Refused to clear the route from a desk. Public network screening is screening, not route clearance. We held that position until exact structures were identified, checked a 4.5 m clearance against the 4.3 m maximum operating height for the movement, and kept the plan at one box per truck until the operator confirmed the geometry.

What you hold at the end

A port decision with the numbers under it — every option priced all the way to your site, not to the wharf.
A container strategy that says what the box is for. A leased asset has a clock. Project material does not. That decision belongs at the term sheet, not the tender.
Free time and equipment terms surfaced as selection criteria, alongside the rate, where they can actually change the answer.
Route constraints closed out or listed open, by name — what was verified, what was only screened, and what is still unresolved. We do not sign off the ones we have not seen.

The finding

We specified a container fleet and it worked. Then we modelled it back on published tariffs, and about four fifths of the exposure had already gone by the time the discharge port moved 900 km closer to site — a change that cost nothing, because it is a line in the purchase order rather than a capital item. Owning the boxes cleaned up what was left. The expensive intervention was not the one that did the work. If you only ever fix one thing on an import programme, fix where it lands.

The detail

The three ports as specified

Northern direct call, about 200 km from the estate. Preferred on cost and transit, and the lane that was executed.
Second northern port, the same class of road leg, held as the alternative where vessel rotation suited better. Constrained by carrier coverage: only one line was found calling there.
Southern port with roughly 1,100 km of inland on-carriage, road or rail. Retained as the contractual backup and the only lane where rail was priced as an alternative to road. It was never activated.

Road constraints, closed and open

A structure reported at 4.5 m clearance, checked against a 4.3 m maximum operating height for the movement. Standard high-cube equipment clears.
A gross mass envelope with a surcharge threshold below the general ceiling, and a tighter limit again at the northern operator. Gross weight per container was requested from the shipper and remained outstanding.
One container per truck as a planning rule, with twin configurations withheld pending operator confirmation of route and access.
An operating-hour restriction on a second structure, raised and still open.

How the detention exposure was estimated

The clock runs from vessel discharge to empty de-hire. On the forced cycle that is three to four days from discharge to gate-out, two days on the road out, a day to unpack, two days dead-hauling the empty back and about a day for the de-hire slot — then nought to five days of queue, depending where the box sat in the vessel block against a site that received four a day.
That is ten to thirteen days under the clock. Against the free-time windows the carriers quoted on this lane it costs nothing at seventeen days, a few tens of dollars a container at fourteen, and roughly $190 to $450 at twelve and ten — with the last boxes off each block reaching about $1,000.
Costed on published Australian tariffs: about $180 per container per day for the first week past free time, rising to $300–$397 a day beyond two weeks (Hapag-Lloyd Australia effective 1 October 2025; ONE Australia valid from 1 April 2026), ex-GST. These are published tariff rates rather than negotiated ones, and an account at forty to fifty boxes a month would contract its own free time and pay less.
The 33-container first wave carried roughly $4,000 to $20,000 of exposure, driven almost entirely by which carrier’s free time applied. Annualised at the intended run rate, and conditional on that rate being reached on the unchanged cycle with no change of carrier, the whole redesign is worth something like $80,000 to $105,000 a year — nil if every box rode a seventeen-day service, and up to roughly $270,000 if every box rode a ten-day one.
Under $25,000 a year of that is attributable to owning the containers rather than to moving the discharge port. We publish the split because the split is the finding.
No detention was invoiced on this programme. Seven containers have shipped, all shipper-owned. These are modelled exposure figures, not amounts anyone paid.

Designing the over-dimension requirement out

For the plant the client wanted to move, we set out what would trigger low-loader configurations, pilot escorts and alternative routing, and the cargo data needed to settle each one.
Then we asked the better question: does any of this have to travel over-dimension at all? Standardising the programme onto a single container type took the escorts, the permits and the route restrictions out of the design.
The cheapest over-dimension movement is the one you engineer away.

Want the long version? The full case study carries the consignment detail, the chain as assembled and the questions we put to each vendor.

A modular housing import programme, 2025 to 2026, four Chinese load ports into North Queensland. The comparative percentage is from the route matrix issued to the client. The international leg was carried by a licensed forwarder.

Which of these is your problem?

The freight was never the problem.

Two routes to market, one pallet

Landed cost per unit, indexed

The line that decided it

“Dangerous goods handling — per unit”

If a unit is a carton

If a unit is a bottle

What we did

What you hold at the end

The detail

The consignment as specified

The chain as assembled

What we would ask the vendors before committing

The site could hit the volume or hit the KPI. Never both.

The service standard, against the site’s own physics

Parcels per day

What actually bound the site, in the order it bound it

A worked example of the same problem, from the same account

“Are you able to provide reverse logistics?”

What we asked back, the same day

What those two answers decide

What we did

What you hold at the end

The detail

The daily cycle

Why the two processing modes mattered

What we would put in the specification next time

Nobody was going to fix this with a rate.

Three discharge ports, one destination

Where the boxes could land, and what each choice cost in distance

The container cycle, before and after

What the clock actually cost

Days the box was held, against the free time each carrier allowed

Where the saving actually came from

Annual detention exposure, at the intended run rate

What we did

What you hold at the end

The detail

The three ports as specified

Road constraints, closed and open

How the detention exposure was estimated

Designing the over-dimension requirement out

Notice that two of these end with the freight not moving.

What you are buying

Who holds what

Tell us the decision in front of you.