# Descartes *The network you can't rebuild* --- Descartes sells logistics software. Underneath, it is the Global Logistics Network, built over decades: a web of electronic connections between the parties that move goods, shippers, carriers, freight forwarders, customs brokers, and the government agencies that regulate all of them. More than two hundred thousand of them are plugged in, across a hundred and sixty countries, and the network carries north of twenty billion messages a year. The value is in the connections, not the software. A carrier joins because that is where the shippers are; a shipper joins because that is where the carriers are; and each one that joins makes the network worth slightly more to everyone already on it. A competitor can write better software in a year. It cannot assemble two hundred thousand trading partners already talking to each other. That is the moat: not the code, the connections, and the connections are the one thing that cannot be rebuilt from scratch. --- Then there is the tailwind, and it is the best kind, because someone else creates the demand. Moving goods across a border means filing data with the authorities, electronically, in the format they demand, or the shipment does not move. Every time a government tightens its rules, the EU's ICS2 cargo-security programme is a recent example, a fresh wave of mandatory filings appears, and it has to flow through somebody's pipes. The incumbent network is where it flows. Descartes never has to persuade anyone to want compliance; the law does the selling, and each new rule adds traffic to the network it already owns. On top of that, Descartes buys. Sixty-odd small acquisitions since 2006, customs-content firms, carrier-connectivity tools, data feeds, each one a new node or capability that the existing network immediately distributes to everyone on it. It is the reverse of the leave-alone acquirers in [[The thin centre]]. The point is not cheap earnings; the network makes each acquired piece worth more than Descartes paid, because it plugs into two hundred thousand parties on day one. --- None of this was ever guaranteed. In 2004, after the dot-com collapse, Descartes was days from failing. It cut roughly a third of its staff, replaced its leadership, and spent years rebuilding around the network instead of chasing growth. The lesson sits in the near-death as much as in the recovery: a network business is worth little until it reaches critical mass, and getting there can nearly kill you. Descartes crossed that line and now throws off cash on around seven hundred million dollars of revenue, 93% of it recurring. It is not immune. The network carries trade, so when trade slows or tariffs bite the traffic thins, and the shares can fall hard, as they did over the past year. Growth leans on a steady supply of acquisitions, which a sceptic rightly tracks. But the core has the quality every operator wants and few can build: demand it does not have to create, and a position a competitor cannot copy without first assembling the very thing that took decades to assemble. ---