Why Rail Freight Needs More Than a CRM
A single unit train on South Africa's Sishen-Saldanha ore line can run to 375 wagons and move over 20,000 tonnes in one working. That number matters, because it's also where most rail software falls apart.
South Africa's freight rail network is opening. In May 2026, Transnet's Rail Infrastructure Manager signed access agreements with 11 private train operators, taking the number of companies moving freight on the national network from one to twelve almost overnight. More will follow as the reform programme matures. For most of these new entrants, one of the first serious decisions, right after the access agreement and before the first train runs, is what system the business gets built on.
Almost everyone reaches for a CRM first. It's the obvious category. Track customers, manage the pipeline, log the contracts. The trouble is a CRM was built to help a salesperson close a deal. It has no concept of a wagon.
To be fair to the category, a CRM earns its place. Rail freight is still, underneath the tonnage, a relationship business: knowing who your customer is, what they've committed to move, on what corridor, in what delivery window, at what rate. That's a real layer, and it's the one every off-the-shelf system handles well. Lead, deal, contract, done. This layer isn't unnecessary. The problem is that people stop there, because it's the layer the software category happens to be named after.
The layers underneath
A signed contract says a customer will move, say, 5,000 tonnes of manganese from a siding in the Northern Cape to a port terminal within a six-week window. Everything that happens between that signature and the invoice is where a CRM goes quiet.
Operations is the first gap. Which wagons, suited to that specific product, get allocated to the contract. Whether they've been placed at the siding for loading, or they're still three days out. Where the consist is right now, mid-corridor. Whether it's arrived and been offloaded, or is sitting in a yard somewhere, unbilled. A CRM can hold a record that says "Deal: Won." It has no way of knowing whether a single wagon behind that deal has moved.
Verification is the second, and it's the layer that protects the number on the invoice. What did the wagon weigh, loaded, at origin, and again at destination, after transit losses, moisture, and handling have all had their say. Does the net weight reconcile against what the contract said would move. And when it doesn't, when a weighbridge reading looks physically implausible, or an identifying tag drops out mid-corridor, does the system flag it as an exception, or does it just silently fail to update, leaving the discrepancy to surface three weeks later at reconciliation, if it surfaces at all.
That last failure mode is the expensive one. Rarely fraud, more often a gap nobody's watching, on a corridor moving tens of thousands of tonnes a week, where a small, unflagged discrepancy per wagon adds up fast across a full unit train.
Established operators have spent decades papering over this gap with people: a foreman who knows the loading yard, a clerk who chases down a missing weighbridge slip, institutional memory that catches most exceptions before they become write-offs. It's not efficient, but it works, mostly, because the workaround has had thirty years to mature. None of that exists for an operator standing up a new service on a corridor it's never run. The eleven companies now onboarding to Transnet's network are building operations from a much shorter runway, on routes carrying coal, manganese, containers, and general freight they haven't moved before. A CRM will tell them who their customers are. It won't tell them whether the physical movement behind those contracts is reconciling, and on a per-tonne business, that's the number that decides whether the margin survives contact with reality.
What we've built instead
This is the thinking behind how we've approached C-WIP: three layers (commercial, operations, and verification) working off one shared record of the physical asset, rather than three separate systems reconciled by hand after the fact. In-motion weighing and automated wagon identification sit underneath the operational layer, feeding it real numbers instead of assumptions, so that "Deal: Won" and "5,000 tonnes actually moved and reconciled" are the same fact, not two records that drift apart and get argued about at month-end.
ECON builds corridor infrastructure for Southern African bulk freight, including C-WIP, its Weighing-as-a-Service platform. Get in touch to talk through what this looks like for your corridor.