Dangerous Goods Shipping Software: Why Your Checks Fail Every January

Why dangerous goods breaks differently from every other compliance problem

It is 4pm at a chemical plant shipping dock. A pallet is built, the truck is booked to a forwarder for an air move to Singapore, and the paperwork is being assembled from a template that a previous safety adviser saved in 2021. The material is a flammable liquid that the plant ships constantly by road, where the rules are familiar and nobody thinks twice. By air it is a different quantity limit, a different packing instruction, and a passenger aircraft prohibition that nobody checked because the road version has been fine for four years. The forwarder catches it at the ramp, or does not. If they do, the shipment misses the flight and the customer misses a production run. If they do not, it becomes an incident report and the carrier stops accepting your consignments while they audit you.

Dangerous goods is not like other compliance work, because the answer changes with the mode, the route, the packaging, the quantity, the concentration and the calendar. The IATA Dangerous Goods Regulations publish a new edition every year and the IMDG Code amends on its own cycle, while the US hazardous materials regulations in 49 CFR sit alongside ADR in Europe. Then individual states and operators publish variations on top, so a shipment legal on one carrier is refused on another out of the same airport.

The operational failure is almost never ignorance of the rules. The failure is that the rules are applied at the end, by one person, to a shipment commercial already promised, on a product whose specification changed nine months ago. By then the only options are delay it or wave it through, and that decision under time pressure is the actual risk here. It is a systems problem, not a training problem.

Problem 1: the substance master is not the product master

Your ERP (Enterprise Resource Planning) knows a material number, a description and a weight. Dangerous goods needs a UN number, proper shipping name, class and subsidiary risk, packing group, flashpoint, marine pollutant status, concentration and whether the packaging in use is authorised. Those attributes live somewhere else, usually in the safety data sheet library or a spreadsheet, keyed by product name rather than by the material number your orders reference.

Labelmaster DGIS, Exis Technologies Hazcheck, IATA DG AutoCheck and DGOffice all carry maintained regulatory content and they are genuinely good at what they do. Their design point is validating or building a declaration for a consignment presented to them. What none of them can do is own your substance master, because your substance master is tied to your formulations, your grades, your concentrations and your packaging specifications. So there is a mapping step, and the mapping step is a human, and the human is why the flammable liquid was still classified against last year's formulation.

What a custom build does: make the dangerous goods attributes a first class part of the material record, versioned, with an effective date, owned by the safety function but visible to everyone. When a formulation changes concentration, the classification review is triggered as a task rather than discovered on a ramp. Machine assistance does one specific job here and it is worth having: safety data sheets arrive as PDFs from suppliers in every layout imaginable, and an extraction pass reads section 14 transport information, compares it to what you currently hold for that material, and raises a difference for a human to approve. It does not classify anything. It notices that a supplier quietly changed something.

Problem 2: the check runs after the promise is made

Order entry takes a line. Planning allocates stock. Shipping builds a pallet. Somewhere near the end, dangerous goods paperwork gets produced, and only then does anyone discover that the quantity exceeds the limit for the intended aircraft, or that the destination state has a variation, or that the customer address is a residential one that a carrier will not serve for this class.

External validation tools reinforce this by design, because they check a consignment that already exists. That is the right place for a final gate and the wrong place for the only gate. The cheapest dangerous goods problem is the one caught in the order line, when a different pack size, mode or split shipment is still available.

What a custom build does: run determination at order entry against the mode and route being quoted. The result is not a document, it is a decision with reasons: this line is acceptable as cargo aircraft only, this quantity would need a different packing instruction, this destination requires an operator variation the chosen carrier does not accept. The salesperson sees it before they commit a date. In our experience this single change removes most of the 4pm arguments, because the argument has already happened at 10am with better options available.

Problem 3: segregation is a consignment property, not a line property

Any tool will tell you whether a single substance is shippable. The harder question is what happens when you put several of them on one pallet, in one container, in one truck. Sea freight has explicit segregation requirements between classes. Road has its own load compatibility rules and quantity thresholds that change the placarding and the driver qualification. Air has state and operator variations that interact with what else is on the aircraft in ways you do not control.

This is where a checking tool and a spreadsheet both stop being useful, because segregation is a property of a combination that only exists at the moment of load building, which happens in your warehouse system. So it gets enforced by supervisors who know the common pairs and are surprised by the rare ones.

What a custom build does: evaluate segregation at the point the load is being built, in the tool the warehouse already uses, with the pallet or container as the unit of evaluation. Blocked combinations are refused before the pallet is wrapped rather than after. The same engine handles the aggregate thresholds that flip a load into a different regulatory category, which is the failure mode that catches people out: four small consignments each individually below a limit, consolidated onto one truck, collectively above it.

Problem 4: the rule editions move and your logic does not

Every January there is a new IATA edition. The IMDG Code amends. Packing instructions get renumbered, quantity limits shift, entries get added. If your logic is a spreadsheet, someone has to reread the changes and update it, and that person is doing this on top of their real job. If your logic is a purchased tool, the vendor updates the content and you inherit it, which is exactly why you should not build a rules library.

We will state this plainly because it is the most expensive mistake available in this domain: do not build the regulatory content. Licence it. Building your own copy of the dangerous goods rules means owning an annual maintenance obligation forever, with legal consequences when you fall behind, and it will not be cheaper. The build is the workflow around the content: your substance master, your order entry, your load building, your evidence trail. The content comes from a maintained source through an interface.

What a custom build does: treat the rules engine as a replaceable dependency behind your own interface, so an edition change is a content update rather than a code change, and so you can move suppliers later without rewriting your operation. It also means every determination is stored with the edition and rule version that produced it, which is what you need when someone asks two years later why a shipment went out the way it did.

Problem 5: proving competence when something goes wrong

After an incident the questions are consistent. Who classified this, against which edition, were they trained and current, what packaging was specified and was it marked to the right standard, who packed it and who signed the declaration. If the answers live across an email chain, a training spreadsheet and a filing cabinet, the finding is usually about your system rather than the shipment.

What a custom build does: bind competence to action. The system will not let an untrained or expired user sign a declaration, because it knows their training record. Every determination is stored immutably with its inputs, the rule version, the user and the timestamp. Packaging specifications carry the UN certification detail and the certificate expiry. When the investigator arrives, the answer is a query rather than a search.

What this costs and how long it takes

Across the 2,000-plus projects Digital Heroes has delivered, this category prices as follows. A focused first release, meaning a versioned substance master, mode-aware determination surfaced at order entry, integration to a licensed rules source, and declaration and document output, runs $80,000 to $170,000 and ships in 14 to 20 weeks. A full platform adding segregation and load compatibility at pallet and container level, multi-leg and multi-mode consignments, label and marking generation, training and competence binding, and a full audit trail runs $200,000 to $500,000 phased over 8 to 14 months.

What pushes the number up: the number of modes, since air, sea and road each carry their own document set. Multi-leg door to door movements, because a consignment that flies then drives must satisfy both regimes and the strictest constraint governs. The size and messiness of your substance master, especially where one commercial product has different classifications by concentration or pack size. And label generation, since compliant marks at the right size and durability are fiddly physical work.

Build versus buy, and when buying wins

Buy if you ship a limited set of UN numbers by road only, from one or two sites, at volumes where a trained person can genuinely check every consignment. A validated commercial tool plus proper training is the correct answer and any build would be indulgent. Buy also if you are a forwarder whose role is checking other people's declarations rather than originating them. That is precisely the job the incumbent tools were designed for.

Build when two or more of these are true. You ship across more than one mode and the same material behaves differently by mode. Your substance data does not live with your orders, so a mapping step exists. You consolidate mixed loads where segregation and aggregate thresholds matter. Multiple sites apply the rules inconsistently. Or a carrier or regulator has already raised a finding, which makes this a licence to operate question.

Our position: the rules are not your intellectual property and you should never own them. The workflow is entirely yours and no vendor will ever fit it, because it is shaped by your products, your order process and your warehouse. Build the second, licence the first, and be suspicious of anyone selling you both as one thing.

How to choose a developer for dangerous goods work

Ask whether they intend to build the rules content. If the answer is yes, end the conversation. The correct answer is an integration to a maintained source behind an interface you own, so the content supplier is replaceable.

Ask how a determination is stored. It must be immutable and must record the rule edition, the inputs, the user and the time. If they describe a status field on the order, they are building something that cannot survive an investigation.

Ask them to explain segregation without prompting. If they treat compliance as a per-line check, they have not understood that the risk is created when lines are combined, and your consolidated loads will be wrong.

Ask who owns the code before kickoff. You should hold the repository, the infrastructure and the freedom to change supplier. At Digital Heroes the client owns the code from the first commit, which here also means you keep the evidence trail if the relationship ends.