Five-Year Cost of Ownership - How Finance Should Compare ERP Proposals
By the end of an ERP selection the CFO has three proposals on the desk, and they do not line up. One is a subscription priced per user per month with the first year discounted. One is a perpetual license with annual maintenance and a hosting fee. One bundles implementation services as an estimate "to be refined." Each vendor's summary page shows a different total over a different number of years. The job for finance is to put all three on the same page - the same scope, the same five years, the same cost lines - so the decision rests on what the company will actually pay.
This guide is for the CFO, the controller and the finance analyst building that comparison at a manufacturer or distributor. It covers why five years is the right horizon, the cost lines proposals leave out, a method for normalizing the quotes, what the comparison tells you to negotiate, and why cost is one column in the decision rather than the decision itself. It is the second gate of the approach in our guide to the business case your CFO will sign.
Why five years
Three years flatters a subscription, because the escalators and the user growth have not arrived yet. Ten years flatters a perpetual license, because the upgrade that will be needed by then is not on the page. Five years covers the implementation year, three years of steady running and the first renewal, which is where subscription uplifts, maintenance increases and user true-ups show up. It also matches the horizon most boards use for a capital decision of this size.
The cost lines proposals leave out
Software - Subscription or license by user type (full, limited, shop floor, warehouse, portal), modules and add-ons, extra environments for testing and development, and the escalator or renewal terms the contract actually allows.
Implementation services - The partner's fees, and whether they are fixed with milestones or time and materials with an estimate. Data migration, integrations (EDI, e-commerce, shop-floor systems, warehouse automation, a third-party logistics provider), reports and forms, customizations and their upkeep at every upgrade, testing and training.
Internal time - Process owners' hours, backfill for the people pulled into the program, a program manager, and the change management nobody budgets. Internal time is often the largest number that never appears in a vendor's proposal.
Running costs - Hosting or infrastructure, upgrades (continuous for a subscription, a project every few years for an on-premises system), the support tier the company will really need, third-party applications the vendor's demonstration quietly relied on, and licenses for growth - a new branch, a new plant, an acquisition.
Transition - Running two systems during cutover, a parallel close, retaining legacy data in a form that can be queried, and decommissioning.
The small print - Price indexation, user-count true-ups, module unlocks, premium support, sandbox limits, API and storage limits. Each one is small. Together they decide year five.
How to normalize the proposals
Put every proposal on the same scope - The same users by type, the same modules, the same sites and the same integrations. If a finalist left something out, ask for it in writing rather than estimating it for them.
Convert to annual cash, year one to year five - One row per cost line, one column per year. A perpetual license lands in year one; a subscription is spread; services follow the implementation plan.
Apply the contract's escalators, not the salesperson's - If the agreement allows a 5 percent annual increase, year five is about 22 percent above year one. Use the number the contract permits.
Add the costs that are the same for everyone - Internal time, transition and change management belong in every column. They differ between finalists only when the approach differs, such as a longer implementation or a heavier data conversion.
Separate one-time from recurring - Show the total over five years and the run rate in year five. The second number is what the company will live with after the project is forgotten.
Ask each vendor to confirm the table - A quote reconciliation, in writing, from each finalist. Vendors correct the table quickly when a competitor's column is next to theirs.
Add a growth case - Two more branches, one acquisition, 30 percent more users. The proposal that scales badly shows itself here, before the contract is signed.
What the comparison usually shows
The lowest year-one proposal is often not the lowest at year five. Discounted first-year subscriptions step up at renewal; estimated services become change orders; a small user count grows into true-ups. A perpetual license that looks expensive up front can run cheaply for three years and then need an upgrade project that costs a large share of the original implementation. Customizations are the quiet multiplier - every one of them is paid for again at each upgrade, which is why configuration rather than customization is a finance question as much as a technical one. Our note on why a Tier 2 shortlist is not enough covers how the tiers compare over the same five years.
What the comparison tells you to negotiate
Renewal and escalator caps - A ceiling on annual increases for the full term and a cap at the first renewal.
User true-up terms - How often counts are reconciled, at what price, and whether reductions are allowed.
Module and user pricing locked for the term - So growth is priced today, not at the moment the company needs it.
Services on a fixed fee with milestones - With acceptance criteria written into the statement of work, and a change-order process that requires the company's approval.
Exit terms - Data return in a usable format, the notice period and what happens to the data after termination.
Named resources - The partner's project manager and lead consultants named in the agreement, with a substitution clause.
The negotiation is also the moment for an independent review of the contract and the statement of work; our decision check before signing describes what to look for.
Cost is one column
The five-year cost of ownership sits beside fit, risk and the partner in the final scorecard. A system that fails the business costs far more than the difference between any two proposals, and most of the risks that sink ERP selections have nothing to do with price. Use the comparison to make the cost honest and the negotiation specific, then decide on fit.
Frequently asked questions
Should we compare on three, five or seven years? Five. Three years hides the first renewal; seven years asks finance to forecast a technology market nobody can see that far into.
Is a subscription always more expensive over time? Not always. A subscription includes upgrades and hosting that a perpetual license pays for separately. The honest comparison puts both on the same five-year table with every line filled in.
How large are implementation services compared with the software? Usually larger than the first year of software, and sometimes several times larger once data migration, integrations and internal time are counted. A proposal where services are a small fraction of the software cost has usually left something out.
Should internal costs be in the comparison? Yes, in every column. They are real cash and real capacity, and leaving them out is how programs end up under-resourced.
What about the cost of doing nothing? It belongs in the business case, not in the proposal comparison. The comparison decides between finalists; the business case decides whether to act.
Who builds the comparison, finance or the adviser? Finance owns the numbers. An independent adviser supplies the cost lines vendors leave out, runs the quote reconciliation with each finalist and keeps the scope the same across columns.
If you have proposals on the desk and want the comparison to be honest before you sign, talk to John about your ERP program.
