Insights
Most published answers to this question come from companies selling the software. This one comes from someone who has owned a global ERP and Microsoft Dynamics 365 estate worth over two hundred million pounds, and it includes the costs the vendor quotes leave out.
By Christopher Fagan, PEF Consulting Partners. 3 September 2026.
For a UK mid market business, a serious ERP implementation typically lands between a few hundred thousand pounds and a few million, all in. Small businesses on cloud products can do it for less; large enterprises spend tens or hundreds of millions. If a quote you have received looks dramatically cheaper than that range, it is almost certainly measuring only part of the cost, and the missing part becomes your problem after the contract is signed.
Software licences or subscription are the visible cost, and usually the smallest of the big four. Implementation services, the integrator’s fees for design, build, testing and deployment, typically run one to three times the software cost, and more when the estate is complex. Then come the two categories most budgets understate.
Your own people are a real cost. An implementation needs your best operators seconded into the programme, which means backfilling their day jobs or accepting the operational risk of not doing so. Budgets that treat client effort as free are the budgets that produce exhausted teams and delayed decisions, which then produce delay costs that dwarf the saving.
Data is the other. Extracting, cleansing, mapping and reconciling data from the old estate is regularly the largest single workstream on the programme, and it is the one most often priced as an afterthought. Every migration I have led or rescued was made or broken on data, never on software.
Integration to the systems staying behind. Environments and tooling. Training that goes beyond a webinar. Dual running, because you will pay for the old estate and the new one at the same time for longer than planned. Contingency that is actually allowed to be spent. Post go live support at a level that matches reality, because the weeks after cutover are when the operation is most fragile. And the cost of decisions not taken, because a programme waiting on a client decision burns its full run rate while it waits.
Three choices move ERP cost more than any vendor discount. Scope discipline: every customisation you decline is money saved twice, once in build and once in every future upgrade. Data honesty: starting the data work before the programme, not during it, converts emergencies into tasks. And independent leadership: an implementation partner marks its own homework unless someone on your side of the table is checking the plan, the estimates and the change requests. On the BP programme, resetting the delivery model and supplier commercials saved twenty five million pounds without slowing delivery.
Build the budget from the operating model outward, not from the licence quote upward. Cost the client team and backfill explicitly. Ring fence the data workstream with its own budget and leadership. Hold contingency at programme level, not hidden in workstream padding. And before any of it, test whether you are ready at all, because readiness gaps are cheap to fix before signature and ruinous after.
Take our free ERP Readiness Assessment for an honest view of where you stand, or talk to us before the business case is signed. Independent advice, no software to sell.
Related reading: how long an ERP implementation really takes and why ERP implementations fail.