System Integration ROI: An Honest Way to Calculate It

Updated May 2026. Rewritten with a practitioner framework for calculating system integration ROI: the four cost categories most spreadsheets miss, three honest benefit types, and an AUD worked example.

System integration ROI is one of the most over-promised numbers in enterprise IT. The vendor deck shows 250 percent return in 18 months. The post-implementation review shows breakeven at month 30, if you squint. The gap is not because the vendor lied. It is because the cost categories were incomplete and the benefits were stacked optimistically. Both are fixable.

We are an AI and automation consultancy. We have scoped, built, and rescued integration programmes for clients in healthcare, recruitment, finance, and professional services. The framework below is the one we use to give CFOs a number they can stake their forecast on, rather than a hope. For the wider system integration topic see our piece on system integration best practices and the production failure modes in system integration challenges.

This piece walks through what system integration ROI actually measures, the four cost categories that go into the denominator, the three benefit categories that go into the numerator (kept strictly separate), a worked example with AUD numbers, the common mistakes that inflate the answer, and when ROI is the wrong question to ask. Pricing in this piece is AUD.

What System Integration ROI Actually Measures

System integration ROI is the ratio of net benefit to total investment over a defined time horizon. Three years is the typical horizon for mid-market integration work. The formula at its plainest is (cumulative benefit minus cumulative cost) divided by cumulative cost. A positive ROI says the work earned its keep.

The reason most system integration ROI calculations are wrong is not the formula. It is the inputs. Two specific failures dominate. First, the cost side typically captures build cost and ignores everything else (run, change, opportunity, sinking fund). Second, the benefit side double-counts: it adds up labour saved, plus efficiency gains, plus revenue uplift, without checking whether the same FTE shows up in two categories. Both fail audit at month 18.

The framework we use comes from working backwards from real post-implementation audits. The categories below are exhaustive on cost and conservative on benefit. The output is a number you can defend in a board paper, not a number that wins the procurement decision.

The Four Cost Categories for System Integration ROI

The denominator of system integration ROI is total cost over the time horizon. Four categories:

Build cost. Vendor fees, internal engineering time at fully-loaded rates, hardware, software licences, third-party integration tools, project management, and architectural review. The line most often understated here is internal engineering time. Multiply the developer’s salary by 1.4 to 1.6 to get a fully-loaded rate that includes super, leave, equipment, and overhead. The hourly rates we see on real projects: senior engineer $180 to $250 AUD per hour, integration specialist $150 to $220 AUD, project manager $130 to $190 AUD.

Run cost. Annual recurring infrastructure, licences, monitoring, on-call rotation, vendor support contracts, security audit, and renewals. We typically budget 15 to 25 percent of build cost per year for run, depending on the technology choice. Custom builds skew higher because of maintenance burden. iPaaS platforms skew lower on engineering but higher on licence fees.

Change cost. Communication, training, process redesign, documentation, user acceptance testing time, hypercare period, and dual-running while the old system is decommissioned. This is the line most often missing entirely from system integration ROI calculations. Budget 10 to 25 percent of build cost. Large enterprise programmes regularly hit 30 percent.

Sinking fund for fixes. Integrations break. APIs change. Schemas drift. Vendors deprecate endpoints. Budget 8 to 12 percent of build cost annually for ongoing fixes and small enhancements. The fixed amount surprises clients but matches what we actually spend on maintenance retainers across our integration portfolio.

The Three Benefit Categories (Keep Them Separate)

The numerator of system integration ROI is total benefit. Three categories. Keep them on three separate rows. Do not stack them into one number without explicit reconciliation, because the same headcount or revenue cannot count twice.

Cash benefits. Reduced headcount, retired licences, eliminated third-party services, avoided contractor fees. These show up on the P&L within the financial year. They are the easiest to defend and the easiest to verify. Cash benefits are the only category we treat as full ROI inputs without discounting.

Cash-equivalent benefits with a redeployment plan. Hours saved per week across an existing team that has a documented plan for what those hours become. “The AP team will reclaim 6 hours per week each and we will use that capacity to absorb the merger volume without hiring two additional FTE.” That is a cash-equivalent benefit because the avoided hire is on paper. “The AP team will reclaim 6 hours per week each and we will be more strategic” is not, because nothing changes on the P&L. We discount cash-equivalent benefits by 30 to 50 percent to reflect realisation risk.

Strategic benefits. Improved customer experience, better data for decisions, reduced compliance risk, increased platform optionality for future work. These are real and important. They are also impossible to verify in a system integration ROI spreadsheet. List them. Discuss them at the board level. Do not put them in the ROI number itself. We have seen too many programmes fail audit because strategic benefits propped up an underwhelming cash story.

System Integration ROI: A Worked Example in AUD

A mid-market professional services firm we worked with integrated their CRM (HubSpot), practice management system, accounting (Xero), and HR system over an 8-month programme. Real numbers, lightly anonymised:

• Build cost: $185,000 AUD (vendor engagement $120,000, internal time $52,000 fully-loaded, software and infrastructure $13,000).
• Run cost: $42,000 AUD per year (hosting, monitoring, support retainer).
• Change cost: $28,000 AUD (training, documentation, parallel-running for six weeks).
• Sinking fund: $18,000 AUD per year for fixes and small enhancements.
• Three-year total cost: $185,000 + $28,000 + (3 x $42,000) + (3 x $18,000) = $393,000 AUD.

Benefits over the same three years:

• Cash: retired two SaaS tools at $1,400 AUD per month each, total $100,800 over three years.
• Cash: did not hire the additional finance role planned for year two ($95,000 AUD per year fully loaded x 2 years = $190,000).
• Cash-equivalent: reclaimed 240 hours per month across the operations team, redeployed to client onboarding capacity which absorbed 12 additional engagements without hiring. Conservative cash-equivalent value: $180,000 AUD over three years. After 40 percent discount for realisation risk: $108,000.
• Strategic (not in ROI number): faster month-end close, fewer customer-facing errors, audit trail for ISO 27001. Worth listing in the board paper.

Three-year total benefit: $100,800 + $190,000 + $108,000 = $398,800 AUD. System integration ROI = ($398,800 – $393,000) / $393,000 = 1.5 percent over three years. That is just above breakeven.

This is the answer the CFO needs to hear before signing. The strategic benefits are real and may push the actual outcome higher, but the cash story is thin. The programme still got approved, with the board accepting that strategic benefit was the actual decision driver. That is the right conversation. The wrong conversation is the version where the strategic benefits get monetised and the ROI line shows 95 percent.

Common Mistakes That Inflate System Integration ROI

Patterns we see in vendor decks and internal business cases that make the number wrong:

Counting time savings without a redeployment plan. “We will save 1,200 hours per year” is meaningless unless there is a documented plan for what those hours become. If nobody loses their job and nobody gets a new assignment, the hours disappear into Parkinson’s law. The ROI line should be zero.

Excluding change management. A $200,000 AUD integration that takes six months to roll out and is unused at month nine has destroyed value, not created it. Change cost is real and it goes in the denominator. Budget it explicitly.

Using book value rates for internal staff. If you use a developer’s base salary divided by 2,000 hours as the hourly rate, you have understated the cost by 40 to 60 percent. Use fully-loaded rates.

Pretending there is no sinking fund. Integrations break. Vendors deprecate. Schemas drift. The maintenance line is non-optional. We have seen $400,000 AUD integration projects with $0 budgeted for ongoing fixes. Two years later they are unmaintained and being replaced. The original ROI was illusory.

Stacking strategic and cash benefits into the same line. The same FTE cannot count toward cost savings and toward strategic uplift. Keep the three benefit categories on separate rows.

The Formula That Survives a CFO Conversation

The system integration ROI formula we use:

3-Year System Integration ROI =
  (Cash Benefits + Discounted Cash-Equivalent Benefits)
  ---------------------------------------------------------
  Build + (3 x Run) + Change + (3 x Sinking Fund)
  - 1

Discount rate on cash-equivalent: 30% to 50%.
Strategic benefits: listed separately, not in this number.

This gets you to a number that survives a six-month post-implementation audit. Anything above 20 percent over three years is genuinely good. Anything above 50 percent should make you go back and check your assumptions. Most ROI numbers above 100 percent are wrong on the inputs, not the formula.

For longer programmes, calculate the payback period as well: cumulative net benefit by month. The payback chart is what most CFOs actually read. ROI is a summary number; payback shows when cash starts coming home.

When System Integration ROI Is the Wrong Question

Honest section. Three integration scenarios where you should not put system integration ROI on the lead slide:

Compliance-driven integrations. If you have to integrate the HR system into the access control system because the auditor said so, the ROI is “we keep our licence to operate.” Trying to calculate a percentage return is forced and unhelpful. Frame it as risk avoidance instead, with the cost of the alternative (lost certification, regulatory action) as the comparator.

Customer experience integrations. If the goal is to give customers a unified view across products, the benefit is retention and lifetime value. These take 18 to 36 months to show up cleanly in the numbers. Putting a three-year ROI on the business case forces a fight you cannot win at month six.

Platform work that enables a strategy. Some integration work is the precondition for the next two years of initiatives. Trying to ROI-justify the precondition in isolation misses the point. Bundle it with the initiative ROI and assess the package.

Frequently Asked Questions

How do you calculate system integration ROI?

System integration ROI is the ratio of cumulative net benefit to cumulative cost over a chosen time horizon, typically three years. Add up four cost categories (build, run, change, sinking fund) and three benefit categories (cash, discounted cash-equivalent, and strategic kept separate). The formula is (cash plus discounted cash-equivalent) divided by total cost, minus one.

What is a good system integration ROI?

Above 20 percent over three years is genuinely good. 50 to 80 percent is exceptional and usually only happens when the integration replaces several expensive licences or removes a large amount of manual handling. Anything above 100 percent should be re-checked; the inputs are usually wrong.

What is the typical payback period for system integration?

Payback periods we see in production: 8 to 14 months for AP automation and invoice processing, 14 to 24 months for CRM and practice management integrations, 18 to 36 months for ERP integrations, and 36 months plus for enterprise-wide platform work. Faster paybacks usually mean the cost side was understated.

Should strategic benefits be in the system integration ROI number?

No. List them separately in the board paper. Strategic benefits are real but unverifiable in a finite spreadsheet, and including them in the ROI line is the single biggest reason integration programmes fail post-implementation audit. Keep cash, cash-equivalent, and strategic on three separate rows.

How much does system integration cost in AUD?

Mid-market integration programmes typically range from $80,000 AUD for a point-to-point integration to $500,000 AUD for a multi-system iPaaS programme. Enterprise event-bus and ESB programmes start at $1 million AUD and can reach $5 million or more for complex multi-entity organisations. Add 15 to 25 percent annual run cost and 8 to 12 percent annual sinking fund on top.

What cost categories are most often missed?

Three. Change cost (training, communication, dual-running during cutover) is missing from most business cases. Fully-loaded internal time (super, leave, equipment, overhead) is often calculated at base salary rates, which understates by 40 to 60 percent. And the sinking fund for ongoing fixes is frequently set to zero, which makes the three-year cost look smaller than it really is.

Does iPaaS or custom integration give better ROI?

iPaaS (Workato, MuleSoft, Boomi, Tray.io, n8n) typically lowers build cost and raises run cost via licences. Custom Python or Node lowers run cost and raises build cost. The crossover varies by integration count: under 10 active integrations, iPaaS usually wins on total cost of ownership. Above 30 active integrations, custom or self-hosted open source often wins.

How do you track system integration ROI after go-live?

Monthly tracking against the baseline established before the integration. The three measurements that matter: actual cash savings against forecast, actual hours saved against forecast (with documented redeployment), and unplanned cost (vendor change orders, fix cycles, missed deprecations). Most programmes that fail post-implementation audit fail because nobody owned this measurement.

If you have a system integration proposal in front of you and the ROI number looks too good, we will work through your inputs with you. Get in touch and bring the spreadsheet. Book a call for a quick review of the cost and benefit categories.