The enshittification of manufacturing software.

You feel it before you can name it. The maintenance invoice goes up every year while the product stands still. The upgrade you were promised turns out to be a re-platform you pay for twice. A capability you already had moves behind a new module with a new SKU. Support was fine until the vendor got acquired, and now the queue is a bot that says it understands and then closes your ticket. Nothing is broken, exactly. Everything is just a little worse than last year, and a little worse than that the year before.

There is a word for it now

Cory Doctorow called it enshittification, and it was mainstream enough to be named Word of the Year on two continents. His version describes consumer platforms. They treat you well to win you, then degrade your experience to serve their real customers, then squeeze those customers to serve their shareholders, until nothing is left. Manufacturing software is not a social network, so the exact model does not transfer. But the engine does, and it is brutally simple. Once leaving is expensive enough, degrading the product becomes profitable. Quality stops being a selling point and turns into a cost to cut.

Manufacturers are the perfect victims

Here is the uncomfortable part. Nobody on earth is more locked in than a manufacturer running a mature MES. You have a decade of validated workflows, genealogy records, and integrations wired into every machine and every ERP call. Your workforce is trained on it. In a regulated plant, replacing it means revalidating product that is already in market. The switching cost is the highest in all of enterprise software, which makes the question "where are you going to go?" more rhetorical for you than for almost any other buyer alive. The harder you are to lose, the safer it is to disappoint you.

The playbook, on your floor

So the pattern runs, and you can watch it happen without leaving the plant. The perpetual license you bought becomes a subscription you rent. The version upgrade is mandatory, and is really a full re-platform. The maintenance line climbs while the roadmap serves the vendor instead of you. And increasingly the vendor is not even the company you chose. Private equity has learned that a locked-in install base is a nearly perfect asset. Acquire it, raise the prices, cut the research, and collect, because the customers cannot leave. The product you depend on becomes a cash flow to be optimized. You are the cash.

Perpetual, defined

A few years ago I watched this happen in slow motion. A customer had bought a perpetual license, the kind you pay for once and own. Then the vendor, a major automation platform, decided the future was subscriptions. The customer, reasonably, asked them to put in writing what "perpetual" actually meant, so everyone was working from the same definition. The answer came back that perpetual covered the version they already had, and nothing new.

They were running an older release with known, serious security vulnerabilities. Every one of those holes was patched in the next version. To get the fixes, they had to move to a subscription. They were free to keep their perpetual license and sit on the vulnerable release, unsupported. And the upgrade itself was quoted at around half a million dollars, because the entire frontend had been rebuilt on a framework several major versions newer. Perpetual, it turned out, meant perpetual access to the exact software you could no longer safely run.

And none of it was personal. It was structural. Public markets pay far more for a dollar of recurring subscription revenue than for a dollar of one-time license sales, so a board under pressure to grow does the math and converts. The customer who already paid stops being a relationship and becomes a conversion target. It is the same force that drives the private equity roll-up, just wearing a different suit. In both cases the product is being optimized for the people who own the vendor, not the people who run their plant on it. That is the quiet engine under all of this. When the software answers to shareholders instead of users, degrading your experience is not a flaw in the plan. It is the plan.

The two bad exits

Faced with a system getting slowly worse, most plants choose between two exits, and both are traps. You can accept it, pay more for less every year, and file it under the cost of doing business. Or you can rip it out and start over, which in an MES is the single hardest and riskiest project in the building, a big-bang cutover of a system your entire compliance history lives inside. One exit funds the decay. The other bets the plant. Neither one touches the actual mechanism, which is the lock-in itself.

The cure is a smaller cage, not a nicer landlord

That is the part worth sitting with. The problem is not that your vendor got greedy. Vendors will always be exactly as greedy as their customers' switching costs allow. The problem is the switching cost. So the cure is not a nicer landlord, it is a smaller cage. Own your system of record instead of renting it. Keep your data portable, in interfaces and formats you control, so that moving is a decision rather than a demolition. Modernize the experience without re-platforming the foundation underneath it. The moment leaving is genuinely possible, staying has to be earned again, and a vendor that has to earn your business every year cannot afford to let the product rot.