Ask most manufacturing leaders whether their core systems are failing and they’ll say no, reasonably enough, because nothing has crashed. Orders are still going out, quality records still get filed, the quoting spreadsheet (the one with the seventeen tabs and the macro nobody quite dares to touch) still produces a number at the end. That’s the issue, these systems rarely fail the way software fails in a demo, with an error message and a stack trace. They drift until the gap between “still running” and “still working for the business” is too wide to explain away.
That gap became harder to explain away in October 2025, when the Jaguar Land Rover cyber incident, which the UK’s Cyber Monitoring Centre classified as a Category 3 systemic event, is estimated to have cost the UK economy £1.9 billion and affected more than 5,000 organisations (The Guardian). It showed, at national scale, just how thin the line is between an ageing system and live production, and it’s done more to get system resilience onto board agendas than any consultancy deck ever has.
Five reasons this happens
The business moved on and the system didn’t get the memo
Every legacy system was built around a set of assumptions, order volumes, product mix, regulatory environment, that made sense when it was specified. Those assumptions are rarely revisited, mostly because revisiting them isn’t formally anyone’s job, so the system carries on serving a business that, in every way that matters, stopped existing years ago. One manufacturer we worked with had grown so organically across divisions that nobody could say with any real confidence, how data actually moved between the core ERP system and the desktop spreadsheets still quietly running half the operation.
The knowledge lives in people, not in the system
The person who’s run it longest knows which report to trust and which one silently drops a decimal place, and very little of that is written down, because writing it down was never as urgent as the next production deadline. With close to 50,000 manufacturing roles currently unfilled and apprenticeship starts down nearly 40% (Make UK), that knowledge is often closer to walking out the door than most leadership teams would like to admit.
Integration debt adds up without ever announcing itself
Most legacy systems now talk to a CRM, an e-invoicing platform, a customer portal, an ERP module bolted on three managers ago by someone who has since left and taken the only diagram of how it all connects with them. That’s how moving a single field on an order form ends up needing a week of testing (and a Slack thread nobody wants to be tagged in).
What satisfied an auditor a decade ago doesn’t now
Customers in regulated supply chains increasingly want structured, auditable data rather than a PDF and a signature. Proving a check happened and proving what it found, when (and by whom) are two different capabilities, and in a tender process, only one of them is still standing by the second round of questions.
Nobody is quite responsible for deciding to replace it
IT doesn’t want to inherit a project that has no commercial sponsor attached to it. Operations really doesn’t want the downtime. And finance (reasonably enough) isn’t about to fund a fix for something that isn’t officially broken. So the system persists by default rather than by merit, and from the outside those two things look identical.
Why the pressure is building now
Alongside the cyber pressure, two further forces are making this a harder year to look away from. Around eight in ten manufacturers report that tariffs have affected their business in some form (Make UK, International Trade Trends 2026), and a similar share are facing higher raw material costs (Make UK/BDO survey). The margin that used to quietly absorb an inefficient system has largely gone, and a thin margin has a way of turning “we’ll deal with it eventually” into a board-level question.
Deciding what actually needs to change
None of this is an argument for replacement as the default answer, and it would be a fairly convenient one for us to make if it were. In plenty of cases the better move is extension and integration: fewer workarounds, better-connected systems, visibility where there currently isn’t any. In others, the cost and risk of keeping an ageing system running does genuinely outweigh the disruption of moving on. Which one applies rarely announces itself without a proper look first, and that look is worth having regardless of which way it points.
If any of this sounds familiar, chances are you already half-know it, in the way most leaders half-know whether a system is still earning its place or just the one nobody’s gotten around to replacing yet. The easiest starting point is usually a straightforward conversation about which one is actually true for you, rather than a problem to work out alone at a desk. We’ve written more on how that assessment plays out in manufacturing specifically, worth a read.
