What Decides an ERP Program's Five-Year Cost

A five-year ERP program typically costs 3 to 5 times first-year software, ERP Research reports, with implementation alone running 1 to 3 times that first-year figure. Where a given program lands inside that range is set mostly during selection, by three decisions made before the systems integrator has written a line of configuration: how tightly customization is scoped, which SI is chosen and on what fee structure, and how much design discipline the requirements register carries into the build.

Customization scope is set in the requirements register, not in the build

Holding customization to under 10 to 15% of requirements saves 20 to 40% of implementation cost, ERP Research reports. That saving is available at exactly one point: when the requirements register is being written and tagged, before any vendor has quoted against it. A register that lists every function the business performs, rather than the lines that actually differentiate one vendor from another, is the single most common way this saving gets lost before the SI is even picked.

The SI is 40 to 60% of the bill, and the selection is where it gets picked

SI fees run 40 to 60% of total implementation cost, ERP Research reports, which makes the systems integrator the largest single line in most programs. A selection that scores the SI on its own criteria, separately from the software, and checks its proposed staffing against named people rather than a generic role list, is making the single decision with the largest effect on the whole five-year number. A selection that treats the SI as an afterthought to the software choice is not.

Key findings

Software is the smallest of the three major cost lines over five years. Implementation (1x to 3x first-year software) and the SI fee inside it (40 to 60% of implementation cost) are both larger, and both are substantially set during selection.
The customization decision has the widest swing of any single lever: holding it under 10 to 15% of requirements saves 20 to 40% of implementation cost, and that decision is made in the requirements register, not in the build.
Design, sometimes called the fit-gap phase, is only 10 to 20% of implementation hours, but it is where the customization and SI-scope decisions above actually get locked in. A rushed design phase tends to produce the custom-code count a tight requirements register was supposed to prevent.
All of this is a planning range, not a quote. The percentages scale with a negotiated software discount; the discipline in the requirements register does not change if the discount does.

Design phase discipline is where the first two decisions either hold or slip

Design is a small share of hours, 10 to 20% of the total, ERP Research reports, but it is where a tight requirements register either survives contact with the build or does not. A fit-gap conversation that reopens "why not just customize this" for every line the register tagged as standard is a design phase quietly re-inflating the customization share the selection worked to hold down.

Frequently asked questions

Does a selection advisor cost more than it saves? There is no single published figure for that comparison; it depends on program size and how far the register, the SI scoring, and the design-phase discipline would have drifted without an outside hand. The lever with the clearest published size is customization scope: holding it under 10 to 15% of requirements saves 20 to 40% of implementation cost, a range usually larger than the cost of the advisory work itself on a mid-market or larger program.

Is the SI choice really as important as the software choice? On cost, it is often more important: SI fees run 40 to 60% of implementation cost, larger than the software line in most programs. Software fit still decides whether the system works for the business; the SI choice decides a larger share of what it costs to get there.

Can an internal team get the same result without outside help? Yes, with the discipline to hold the requirements register to differentiating lines, score the SI separately from the software, and keep the design-phase fit-gap conversation tied back to the register. An ERP selection advisory engagement exists for teams who want that discipline enforced by someone without a stake in the outcome, not because an internal team is incapable of it.

How does this connect to program risk later? A customization share that drifted up during design and build is one of the inputs behind a program showing the warning signs covered in when an ERP program needs an outside review; cost and schedule risk usually show up together, not separately.

What decides an ERP program's five-year cost

Customization scope is set in the requirements register, not in the build

The SI is 40 to 60% of the bill, and the selection is where it gets picked

Design phase discipline is where the first two decisions either hold or slip

You have read your free articles this month.