Where ERP Return on Investment Actually Comes From
When a CFO or IT manager takes an ERP proposal to the board, the questions never change: what will it cost, and when do we get our money back? The usual answers do not persuade anyone. "It will improve efficiency." "We will get better visibility." "Companies that implemented it saw great results." All true in a general sense, none of it a basis for a decision worth hundreds of thousands of riyals.
ERP does not generate a return by itself. The return comes from the changes it makes possible: less inventory because planning is based on accurate balances, less time spent re-keying and reconciling because data is entered once, faster collection because invoices go out on time, and fewer errors because controls are built into the process. If the way people work does not change, the return does not arrive, however good the software is.
This guide shows how to build an ROI case that a board and an auditor will accept: which costs to include in full, which benefits are measurable and how to measure them, and a five-year worked example using three measures, simple ROI, payback period and net present value. It closes with why returns go missing and how to track them after go-live. All figures in the examples are hypothetical.
The Basic Formula
ROI = (total benefits − total costs) ÷ total costs × 100
The formula is easy; the inputs are where cases fall apart. Two errors spoil most business cases: counting only part of the cost, usually the license or subscription, and inflating benefits with generic numbers unrelated to the company's operations. So we start with cost.
Total Cost: What Belongs in the Calculation
| Cost item | Nature | Notes |
|---|---|---|
| Subscription or licenses | Recurring or one-off | Depends on users or modules |
| Implementation and setup | One-off | Process analysis, chart of accounts, workflows |
| Data migration | One-off | Items, customers, suppliers, opening balances, and cleaning them |
| Customization and integrations | Mostly one-off | Screens, reports, approvals, links to other systems |
| Training | One-off, plus new joiners | By number of users and roles |
| Internal team time | One-off | The most forgotten item: staff hours in workshops, testing and sign-off |
| Infrastructure | Depends on deployment | Servers and backup on-premise; far less in the cloud |
| Support and updates | Recurring | Included in subscription or under a separate contract |
Internal time deserves attention. If the finance manager, two accountants, the storekeeper and the sales manager spend 15% of their time on the project for four months, that is a real cost even though no invoice arrives for it, because the work they would otherwise have done gets delayed. For more on this, see our article on total cost of ownership. On pricing models, Tranquil's pricing page describes per-user pricing for small firms and a flat per-module fee with unlimited users for companies above 50 users.
Measurable Benefits
Group benefits into three buckets, and only put the first two into the core calculation.
Bucket 1: direct savings
- Lower inventory carrying cost. Accurate balances and reorder points cut excess stock. The saving is not the released stock value itself, it is the cost of carrying it: financing, storage and obsolescence.
- Staff time. Hours of double entry, manual reconciliation and spreadsheet reporting. The saving is only real if that time is redeployed to useful work, or if it avoids a hire the company would otherwise need as it grows.
- Less waste and write-off. Expired stock, count variances, supplier invoices paid without a matching receipt.
- Retired tools. A standalone invoicing tool, a separate payroll package, and the spreadsheets maintained around them.
Bucket 2: cash flow improvement
- Lower days sales outstanding. Invoices issued on delivery, overdue alerts, accurate statements. Each day cut from average collection releases cash.
- Early payment discounts from suppliers, now that obligations are visible and scheduled.
Avoided cost of growth
Often overlooked: if the company plans to double volume in three years, how many extra accountants and data entry staff would it need on current processes? Say three, at a loaded SAR 10,000 a month each, so SAR 360,000 a year. If an integrated system absorbs the growth with one extra hire, the avoided cost is SAR 240,000 a year from the year growth happens. That is not a saving on today's costs; present it as avoided future cost, with the growth assumptions it rests on.
Bucket 3: real but hard to quantify
Better decisions from available data, the ability to open branches without proportional admin headcount, audit and financing readiness, and lower risk of non-compliance with e-invoicing and wage rules. Describe these in the case, but do not put numbers on them, or the whole case loses credibility.
Mapping Benefits to Modules
| Module | Expected benefit | Indicator | How to measure |
|---|---|---|---|
| Inventory | Less stock, fewer stockouts | Average stock value, turnover | Monthly against prior-year average |
| Purchasing | Better prices, accurate payments | Price variance, first-pass match rate | Three-way match reports |
| Sales and collections | Faster cash | DSO | Receivables ÷ credit sales × 365 |
| Accounting | Faster close, fewer errors | Days to close, correcting entries | Monthly close log |
| HR and payroll | Less payroll prep time | Hours per payroll run, corrections | Time logged before and after |
| E-invoicing | Retire a separate tool and double entry | Tool cost, rejected invoices | Previous contracts, platform reports |
Small hypothetical example: four accountants take ten working days to close the month, half of it spent reconciling inventory, sales and the ledger. After integration the close takes four days. That saves 6 days × 4 people = 24 days a month, about 288 a year; at a loaded SAR 450 a day, roughly SAR 130,000. The bigger value may be management getting the month's numbers in week one instead of mid-next-month. We leave this out of the main example below to avoid counting staff time twice.
A Five-Year Worked Example
Hypothetical example: a distributor with SAR 60 million revenue and SAR 8 million average inventory, currently on accounting software plus separate spreadsheets for stock and sales.
Costs (SAR thousands)
- Implementation and data migration: 150
- Customization and integrations: 60
- Training: 30
- Internal team time: 90
- Year 0 total before go-live: 330
- Annual subscription including support: 120 for each of five years, so 600
- Five-year total cost: 930
Annual benefits at steady state (SAR thousands)
- Inventory down 10%, or SAR 800,000, at a 20% carrying cost: 160
- Staff time equal to two people redeployed instead of hiring, at SAR 10,000 a month each: 240
- DSO down five days on SAR 60 million revenue releases about SAR 822,000; at 7% financing cost: 57.5
- Lower write-offs and count variances: 50
- Total annual benefit: 507.5
Year one after go-live is assumed to deliver half the benefit while people learn and data settles. Five-year benefits: 253.75 + 4 × 507.5 = 2,283.75.
| Year | Cost | Benefit | Net cash flow | Cumulative |
|---|---|---|---|---|
| 0 | 330 | 0 | −330 | −330 |
| 1 | 120 | 253.75 | 133.75 | −196.25 |
| 2 | 120 | 507.5 | 387.5 | 191.25 |
| 3 | 120 | 507.5 | 387.5 | 578.75 |
| 4 | 120 | 507.5 | 387.5 | 966.25 |
| 5 | 120 | 507.5 | 387.5 | 1,353.75 |
The three measures
- Simple ROI: (2,283.75 − 930) ÷ 930 = 145.6% over five years.
- Payback: 196.25 remains at the end of year one, and year two delivers 387.5, so payback lands about halfway through year two, roughly 1.5 years after go-live.
- NPV at 10%: discount each year's net cash flow to today and subtract the year-zero investment: about SAR 908,000. Positive NPV means the project more than covers its cost of capital.
We did not count the released cash itself, SAR 800,000 from inventory and SAR 822,000 from receivables, as a benefit. That is a one-off liquidity effect worth stating in the case, but adding it on top of the carrying-cost saving would count the same effect twice.
Sensitivity: What If You Are Too Optimistic?
A good board will ask what happens if half of this does not materialize. Answer before they ask:
| Scenario | Steady-state annual benefit | Five-year net benefit | ROI |
|---|---|---|---|
| Conservative: 60% of expected benefit | 304.5 | 440.25 | 47% |
| Base case | 507.5 | 1,353.75 | 146% |
| Costs 30% higher | 507.5 | 1,074.75 | 89% |
Conservative case: 152.25 + 4 × 304.5 = 1,370.25 of benefit, less 930 of cost, gives 440.25. Higher-cost case: costs of 1,209 against benefits of 2,283.75. If your own conservative case turns negative, the project probably needs a smaller first phase. You can try your own numbers in Tranquil's ROI calculator.
Presenting the Case to the Board
- Today's problem in numbers: DSO, count variances, days to close, manual hours.
- Scope: modules in phase one and what waits.
- Five-year total cost, every line including internal time.
- Measured benefits, each with its assumption, source and owner.
- The three measures: ROI, payback, NPV.
- Sensitivity: conservative and higher-cost cases.
- Risks and mitigation: data cleanup, training, change management.
- Tracking plan: post-go-live review dates.
That structure moves the discussion to assumptions rather than impressions. For the implementation phases the plan will rest on, read our guide to ERP implementation.
Why Returns Go Missing
- Old processes carried over: if paper approvals and shadow spreadsheets survive go-live, staff time does not fall.
- Dirty data: duplicate items and wrong balances destroy trust, and people drift back to Excel.
- Thin training: users who know a fraction of the system never switch on reorder points or overdue alerts.
- Over-customization: higher cost, later go-live, harder upgrades.
- No baseline: without measuring DSO, stock accuracy and close time beforehand, you cannot prove improvement afterwards.
- No benefit owners: inventory savings belong to the supply chain manager, DSO to the credit manager. Unowned benefits tend not to happen.
Tracking ROI After Go-Live
- Record the baseline before starting: average stock, DSO, days to close, count variances, overtime in finance and the warehouse.
- Set review dates at 3, 6 and 12 months after go-live.
- Compare actuals to the case, line by line, and explain gaps.
- Correct course: if stock has not fallen, are reorder points set? If DSO has not moved, are overdue alerts used?
- Update the case yearly, and add modules only once current ones deliver.
For a fuller list of indicators, see our article on ERP implementation KPIs.
The Saudi Angle
Some benefits in the Kingdom are about compliance as much as efficiency. With e-invoicing integration waves extending to businesses above SAR 375,000 in revenue in Wave 24 and above SAR 187,500 in Wave 25 from February 2027, many companies need a compliant solution regardless. When the same system issues the e-invoice, applies 15% VAT and links sales to stock and the ledger, the cost of a separate invoicing tool becomes an avoided cost you can legitimately include. The same logic applies to payroll and wage protection when they run inside the same platform.
Phasing also matters for Saudi mid-market companies. Starting with accounting, e-invoicing, inventory and sales covers the compliance-driven needs and the largest measurable benefits first. HR, payroll, projects or assets can follow once the first phase has delivered against its baseline, which keeps early costs down and gives the board evidence before the next commitment.
Frequently Asked Questions
What is a reasonable ERP payback period?
There is no standard figure, because it depends on company size, scope and implementation quality. What matters is that it rests on your own numbers and survives the conservative scenario, not that it matches a figure from a generic study.
Should released inventory count as a benefit?
Count the annual carrying-cost saving as a recurring benefit and state the released amount as a one-off cash effect. Adding both to benefits double counts.
Is saved staff time a real benefit?
Only if it turns into value: analysis, better customer service, or a hire you avoid as you grow. If the time simply goes unused, the benefit is theoretical.
How does ROI differ between cloud and on-premise?
Cloud is usually lower upfront and slightly higher in recurring cost, so payback is often shorter. On-premise needs upfront servers and licenses and may cost less over a very long horizon in some cases. Model both over at least five years.
Can Tranquil help with the ROI case?
Start with the ROI calculator, then contact the Tranquil team for a quote within 24 hours to anchor the cost side of your case.