Cost Estimates and Selection Checklists
ERP ROI and payback: a worked example with realistic savings
How to estimate ERP efficiency savings a Canadian SMB can defend, build net annual savings, adjust for ramp-up after go-live and calculate payback and five-year return.
An ERP pays back when the net yearly saving it creates, counted only from processes that actually change, adds up to the one-time cost of the project. To estimate it honestly: list savings by process and price each one, subtract any increase in yearly software costs compared with what you pay today, assume savings start slowly after go-live, and then calculate the payback period and the net saving over five years. If the result depends mainly on "better visibility" or "growth", treat it as a strategic decision and say so, rather than dressing it up as a financial one.
This guide is part of the ERP selection checklist and uses the ERP payback calculator. All figures are illustrative.
Where ERP savings usually come from
| Saving | How to measure it | Evidence to collect before the project |
|---|---|---|
| Staff hours on manual work | Hours per week × loaded hourly cost | Time logs for re-keying, reconciliations, spreadsheet consolidation, month-end |
| Inventory carrying cost | Lower average inventory × carrying-cost percentage | Current average inventory, stock-outs, write-offs |
| Fewer errors | Credit notes, shipping errors, duplicate payments avoided | Last 12 months of credits and adjustments |
| Faster collections | Days sales outstanding reduced × cost of financing | Current receivables ageing |
| Retired systems | Old licences, add-ons, server and support no longer paid | Current invoices |
| Avoided hires | Growth handled without an extra clerk | Hiring plan |
Loaded hourly cost means salary plus employer payroll costs (CPP or QPP, EI, QPIP in Quebec, workers' compensation, benefits, vacation) divided by productive hours. A rough rule some finance teams use is to add a percentage to base pay; calculate it from your own payroll records rather than guessing.
Building the net yearly saving (illustrative)
A wholesale distributor in Manitoba with 25 staff estimates:
- Hours saved: the AP clerk saves 6 hours a week on three-way matching, the controller 4 hours a week on month-end consolidation, warehouse staff 8 hours a week on spreadsheet stock counts, and sales support 2 hours a week on order status questions: 20 hours a week × 50 weeks = 1,000 hours. At a loaded $40 an hour: $40,000.
- Inventory: better reorder points reduce average inventory by $60,000; at an illustrative carrying cost of 20% a year (storage, insurance, financing, obsolescence) that is $12,000.
- Software cost change: the new subscription, add-ons and support cost $32,000 a year; the old accounting system, add-ons and server upkeep cost $20,000. The increase is $12,000.
Net yearly saving = $40,000 + $12,000 − $12,000 = $40,000.
Notice what is not counted: faster collections and fewer errors are likely but were not measured, so they are kept as upside rather than included. That is a deliberate choice that makes the estimate easier to defend.
Payback without and with ramp-up
The one-time project cost is $120,000 (implementation, migration, training and internal time). No grant has been approved, so the subsidy is $0. The business assumes savings grow 2% a year with wages.
Simple model (savings from day one): year 1 $40,000; year 2 $40,800; year 3 $41,616; year 4 $42,448; year 5 $43,297. Cumulative after three years: $122,416, which passes $120,000 during year 3; payback ≈ 2 + (120,000 − 80,800) ÷ 41,616 ≈ 2.9 years. Five-year net saving ≈ $208,161 − $120,000 = $88,161. This is what the payback calculator shows with its default inputs.
Ramp-up model (more realistic): savings in year 1 are only half, because go-live happens partway through the year and staff are still learning. Year 1 $20,000; then the same as above. Cumulative: $20,000 → $60,800 → $102,416 → $144,864. Payback ≈ 3 + (120,000 − 102,416) ÷ 42,448 ≈ 3.4 years. Five-year net saving ≈ $188,161 − $120,000 = $68,161.
To approximate ramp-up in the calculator, either reduce the yearly saving or add the missing year-1 saving to the project cost (here, add $20,000 to the cost).
Stress-test the result
| Scenario (illustrative) | Payback | Five-year net saving |
|---|---|---|
| Base case with ramp-up | about 3.4 years | about $68,000 |
| Project costs 25% more ($150,000) | about 4.1 years | about $38,000 |
| Hours saved are 30% lower | about 4.6 years | about $12,000 |
| Both | beyond 5 years | negative over 5 years |
If the decision only works in the base case, the project is fragile. Either reduce the one-time cost (narrower first phase, more internal work), increase confidence in the savings (measure current hours properly), or accept that the case is strategic.
Savings that do not show up as cash
Hours saved only become money if you redeploy people to higher-value work, avoid a hire or reduce overtime. Be explicit in the business case about which of these will happen. Otherwise the payback is on paper only.
FAQ
What is a good ERP payback period?
There is no standard. Many owners compare it with the contract length or system life: payback well inside five years leaves room for things to go wrong.
Should I include grants?
Only grants you have been approved for in writing. Many programs have closed or changed; see government funding and tax treatment.
Do tax deductions change payback?
They affect the after-tax cost. Subscriptions are usually deductible as expenses and purchased software may qualify for capital cost allowance; ask your accountant to run an after-tax version.
How do I measure hours saved before the project?
Ask staff to log time on the target tasks for two to four typical weeks, including month-end.
Sources
- Canada Revenue Agency — employer payroll costs (CPP, EI) for loaded hourly cost
- Revenu Québec — Quebec employer contributions (QPP, QPIP and others)
- Statistics Canada — wage and industry data for benchmarking labour costs
