Deferral is a decision with a running cost. This paper quantifies where that cost accumulates and offers a framework for deciding whether to defer again.
Every mid-market manufacturer we work with has, at some point, decided not to replace its ERP this year. That decision is almost always defensible in isolation and almost never costed.
This paper sets out where the cost of deferral actually accumulates — in stale cost standards, in manual reconciliation, in decisions taken on figures nobody has validated — and offers a framework for making the next deferral deliberately rather than by default. Our position is not that every manufacturer should replace its ERP. It is that the decision should be made with the running cost of the alternative on the table.
If you read nothing else, read these. The analysis that follows sets out the evidence for each.
The visible cost of an ageing ERP is the manual effort around it. The material cost is quoting from a cost basis nobody has validated, which distorts every pricing and portfolio decision the business makes.
Nothing in the management accounts is labelled 'cost of stale standards'. It appears as variance, as margin below expectation, and as customers who are quietly unprofitable.
Actual costing, capacity posting and dimensional reporting are available at a licence cost a hundred-person manufacturer can defend. The barrier is no longer the technology.
A programme scoped to prove itself before the next phase is commissioned has a materially different failure mode from one that produces nothing for a year.
Ask a manufacturing finance director what a unit costs to make and you will get a confident answer. Ask when the standard was last reviewed and the conversation changes.
This is not negligence. Updating standards is disruptive, the variance analysis is uncomfortable, and there is never a good quarter to do it. So the number ages while sales continues to quote from it, and the gap between the standard and reality widens quietly.
The consequence appears in three places, in ascending order of expense. Work is lost that should have been won, because the standard is high. Work is won that should have been lost, because the standard is low. And the business cannot distinguish between the two, because variance is explained after the quarter rather than acted on during it.
The barrier to actual costing was never conceptual. It was that capturing real labour, material and machine consumption at the operation required a system most mid-market manufacturers could not justify.
Business Central now carries production orders, routings, bills of material and cost accounting in the same application as the ledger. Microsoft's documentation sets costing method at item level rather than company level, which means a migration can be staged by item category rather than attempted as a single event — a materially different risk proposition from the one most manufacturers last evaluated.
The 2026 release direction adds agents into payables and expense processing, which matters disproportionately in a small finance team where routine document handling is a large share of the week.
Deferral buys three things, and they are worth naming honestly. It avoids disruption during a period the business cannot absorb it. It preserves capital for something with a clearer return. And it avoids a change programme the leadership team does not currently have the appetite to run.
All three are legitimate. The question this paper puts is whether they are worth the running cost, and whether that comparison has ever been made explicitly.
In our experience the answer varies genuinely by organization. A manufacturer with stable products, long runs and infrequent repricing carries a much lower deferral cost than one with high mix, frequent quoting and volatile input prices. The second should be considerably more anxious about this than the first.
Every paper in this series ends with a framework you can run internally. We would rather you used it and reached your own conclusion than took ours on trust.
Five questions to answer before deferring again. If three or more resolve against you, deferral is costing more than the programme would.
When was the cost standard last reviewed against actual production, and by whom?
What proportion of quotes are priced from that standard, and what revenue does it touch?
Can you identify which customers or product families are unprofitable, and how confident are you?
What does the workaround estate cost annually in effort, and what would phase one cost?
Is deferral a decision this year, or the absence of one? Write down which.
The same argument lands differently across an executive team. These are the three versions worth separating.
Microsoft's own documentation for the product behaviour described above. We would rather you verified the basis than accepted our summary of it.
On these references: each entry names a Microsoft Learn article or documentation area by title, because deep links change while titles are stable. Searching the title on learn.microsoft.com will reach the current version. Where we have cited a figure or a product behaviour, it is Microsoft's statement rather than ours; where we have given a number of our own it is labelled as such in the text.
If you would like the deferral cost calculated against your own figures rather than in the abstract, send us a bill of material, a routing and one closed period. We will construct the comparison and tell you honestly if deferral is the right answer.
Ask a manufacturing executive what a security incident would cost and the answer involves stolen designs. Ask what a week of stopped production would cost and the number is immediate and much larger.
Architectures that assume connectivity fail in exactly the places manufacturers need them most. This paper examines the design decision and the failure nobody plans for.
The parallel quarter is the whole method. Switching without one produces an accurate number nobody trusts.
Describe the situation in your own words.
