ROI & Business Case Writing

Key Dimensions

Metric	Definition	When to Use
ROI	(Net Benefit / Cost) x 100	Simple comparison of investment return
NPV	Sum of discounted future cash flows - initial investment	Multi-year investments, accounts for time value of money
IRR	Discount rate at which NPV = 0	Comparing investments of different sizes
Payback Period	Time to recoup investment	Quick filter – “when do we break even?”
TCO	Total cost over the lifecycle	Build vs buy, vendor comparison

---

Financial Metrics – How to Calculate Each

ROI (Return on Investment)

The simplest metric. Use for quick justification of straightforward investments.

ROI = (Net Benefits - Costs) / Costs x 100

Example: CI/CD Pipeline Investment
  Cost: €80,000 (tooling + 2 engineers for 2 months)
  Annual benefit: €120,000 (faster releases, fewer incidents, less manual work)
  
  Year 1 ROI = (€120,000 - €80,000) / €80,000 x 100 = 50%
  3-Year ROI = (€360,000 - €80,000) / €80,000 x 100 = 350%

Limitation: ROI doesn’t account for when the benefits arrive. €120K today is worth more than €120K in 3 years. For multi-year investments, use NPV.

NPV (Net Present Value)

NPV discounts future cash flows to their present value. A positive NPV means the investment creates value.

NPV = -Initial Investment + Sum (Cash Flow_t / (1 + r)^t)

Where:
  r = discount rate (typically 8-12% for internal projects, ask your finance team)
  t = year number

Example: Platform Modernization
  Initial investment: €400,000
  Annual savings: €200,000 for 4 years
  Discount rate: 10%

  Year 0: -€400,000
  Year 1: €200,000 / 1.10^1  = €181,818
  Year 2: €200,000 / 1.10^2  = €165,289
  Year 3: €200,000 / 1.10^3  = €150,263
  Year 4: €200,000 / 1.10^4  = €136,603
  ─────────────────────────────────────
  NPV = -€400,000 + €633,973 = +€233,973

  NPV is positive → Investment creates value.

Decision rule:

• NPV > 0 → Invest (creates value)
• NPV = 0 → Break even (indifferent)
• NPV < 0 → Don’t invest (destroys value)

IRR (Internal Rate of Return)

IRR is the discount rate that makes NPV = 0. It tells you the “yield” of your investment.

Find r where: 0 = -Initial Investment + Sum (Cash Flow_t / (1 + r)^t)

Using the platform modernization example:
  IRR ≈ 35% (solve numerically or use Excel's IRR function)

Interpretation:
  If your company's hurdle rate (minimum acceptable return) is 12%,
  and your IRR is 35%, this is a strong investment.

When IRR is useful: Comparing investments of different sizes. Project A costs €100K with 40% IRR. Project B costs €1M with 25% IRR. IRR helps compare, but NPV tells you absolute value created (Project B likely wins on NPV).

Payback Period

How long until you break even. The simplest and most intuitive metric for executives.

Simple Payback = Initial Investment / Annual Net Benefit

Example:
  Investment: €200,000
  Annual benefit: €80,000
  Payback: 200,000 / 80,000 = 2.5 years

Discounted Payback (accounts for time value):
  Year 1: €80,000 / 1.10 = €72,727 → Cumulative: €72,727
  Year 2: €80,000 / 1.21 = €66,116 → Cumulative: €138,843
  Year 3: €80,000 / 1.33 = €60,105 → Cumulative: €198,948
  Year 4: €80,000 / 1.46 = €54,641 → Cumulative: €253,589
  
  Discounted payback: ~3 years (vs 2.5 years simple)

CFO rule of thumb: Most CFOs want payback under 2 years for infrastructure investments. Under 18 months is a strong case. Over 3 years is a hard sell unless it’s strategic.

---

Quantifying Engineering Value – The Hard Part

The challenge with engineering business cases is quantifying benefits that feel intangible. Here’s how to make them concrete.

Revenue-Linked Benefits

Benefit	How to Quantify	Example
Faster time to market	Revenue per day of delay x days saved	Feature generates €10K/day -> 30 days faster = €300K
Higher conversion rate	Revenue x conversion improvement x traffic	€50M revenue x 0.5% conversion lift = €250K
Reduced downtime	Revenue per minute x expected minutes saved	€500/min x 60 min fewer outages/year = €30K
New market entry	Projected revenue from new capability	New channel generates €2M in year 1

Cost-Avoidance Benefits

Benefit	How to Quantify	Example
Reduced manual work	Hours saved x fully loaded hourly rate	20 hrs/week x €65/hr x 52 weeks = €67,600
Reduced incidents	Cost per incident x incidents prevented	€5K/incident x 24 fewer incidents = €120K
Avoided hiring	Fully loaded cost of hires not needed	Automation avoids 2 hires = €230K/year
Reduced cloud waste	Current waste x optimization %	€50K/month waste x 40% reduction = €240K/year

Productivity Benefits

Benefit	How to Quantify	Example
Developer velocity	Hours per deploy x deploys x cost/hour	2 hrs saved/deploy x 200 deploys x €65 = €26K
Reduced context switching	Hours recovered x cost/hour x engineers	3 hrs/week x 16 engineers x €65 x 52 = €162K
Onboarding time	Weeks saved x new hires x weekly cost	2 weeks saved x 6 hires x €2,750 = €33K

Risk-Reduction Benefits

These are harder to quantify but important. Use expected value:

Expected Value of Risk Reduction = Probability of Event x Impact of Event

Example: Security Investment
  Probability of data breach: 15% per year (industry average for retail)
  Average cost of breach: €4.5M (IBM Cost of Data Breach Report 2024)
  Expected annual loss: 0.15 x €4.5M = €675K
  
  Investment in security tooling: €200K/year
  Reduces probability to 5%
  New expected loss: 0.05 x €4.5M = €225K
  
  Risk reduction benefit: €675K - €225K = €450K/year
  ROI on security investment: (€450K - €200K) / €200K = 125%

---

Tech Debt ROI – Special Case

Tech debt is one of the hardest business cases to make because the costs are distributed and the benefits feel “just engineering quality.” Here’s how to make it concrete.

The Tech Debt Tax Model

Tech Debt Tax = Additional time spent due to tech debt, expressed as % of team capacity

Measurement:
  Track over 2-4 sprints:
  - Time spent on workarounds
  - Time spent on incidents caused by tech debt
  - Time spent on extra testing because of brittle systems
  - Time waiting due to slow builds/deployments

Example:
  Team of 8 engineers
  20% of time lost to tech debt = 1.6 engineers worth of capacity
  Annual cost: 1.6 x €115,000 = €184,000/year in lost productivity

Business case:
  Investment to reduce tech debt: €200,000 (2 engineers, 3 months)
  Reduces tech debt tax from 20% to 8%
  Annual savings: (20% - 8%) x 8 x €115,000 = €110,400/year
  Payback: 200,000 / 110,400 = 22 months

Tech Debt Categories for Business Cases

Category	Business Impact	Quantification
Architecture debt	Feature velocity slowing, scaling problems	Deployment frequency, lead time trends
Test debt	More bugs in production, slower releases	Bug escape rate, time in QA
Infrastructure debt	More incidents, longer recovery	MTTR, incident frequency
Documentation debt	Longer onboarding, knowledge silos	Onboarding time, bus factor
Dependency debt	Security vulnerabilities, compatibility issues	CVE count, upgrade effort

---

Business Case Template

The One-Page Business Case

This is the format that gets read. CFOs and VPs don’t read 20-page documents. They read one page and ask questions.

BUSINESS CASE — [Project Name]
Date: [Date]  |  Author: [Name]  |  Sponsor: [VP/Director Name]

THE PROBLEM
[2-3 sentences. What's broken? What's the business impact of doing nothing?]

THE PROPOSAL
[2-3 sentences. What are we going to do? At what scope?]

INVESTMENT REQUIRED
  People:           €_________ ([X] engineers for [Y] months)
  Tools/Licenses:   €_________
  Cloud:            €_________
  Other:            €_________
  TOTAL:            €_________

RETURNS (3-Year)
  Revenue impact:         €_________ (how)
  Cost savings:           €_________ (how)
  Risk reduction:         €_________ (how)
  TOTAL 3-YEAR BENEFIT:   €_________

KEY METRICS
  ROI:                    ____%
  NPV (10% discount):    €_________
  Payback period:         ___ months
  IRR:                    ____%

RISKS & MITIGATIONS
  1. [Risk] → [Mitigation]
  2. [Risk] → [Mitigation]
  3. [Risk] → [Mitigation]

ALTERNATIVES CONSIDERED
  1. [Do nothing] — Impact: [what happens]
  2. [Alternative approach] — Why not: [reason]
  3. [This proposal] — Why: [reason]

TIMELINE
  Q1: [milestone]
  Q2: [milestone]
  Q3: [milestone/completion]

DECISION REQUESTED
  Approve €_________ for [project] starting [date]

The Detailed Business Case (For Large Investments >€500K)

For investments above €500K, you’ll typically need a more detailed document. Structure it as:

1. Executive Summary (1 page) – The one-pager above
2. Current State Analysis (1-2 pages) – What’s broken, with data
3. Proposed Solution (2-3 pages) – What we’ll build, architecture, team
4. Financial Analysis (2-3 pages) – Detailed cost model, NPV, sensitivity analysis
5. Risk Analysis (1 page) – Risk register with probability, impact, mitigation
6. Implementation Plan (1-2 pages) – Timeline, milestones, governance
7. Appendix – Supporting data, vendor quotes, benchmarks

---

Presenting to Finance – What They Care About

CFO Mental Model

When reviewing your business case, a CFO is thinking:

1. Is this real? Are the numbers credible, or is this engineering hand-waving?
2. What’s the alternative? What happens if we don’t do this?
3. When do we see returns? Payback period is king.
4. What could go wrong? Sensitivity analysis – what if costs are 30% higher or benefits 30% lower?
5. Is this the best use of capital? Compared to other investments competing for the same budget.

Sensitivity Analysis – Show You’ve Thought About Risk

SENSITIVITY ANALYSIS — Platform Modernization

                        Pessimistic     Base Case     Optimistic
────────────────────────────────────────────────────────────────
Investment              €520K (+30%)    €400K         €360K (-10%)
Annual savings          €140K (-30%)    €200K         €240K (+20%)
Timeline                18 months       12 months     9 months

NPV (10%, 4 years)     +€42K           +€234K        +€385K
Payback                3.7 years       2.0 years     1.5 years
ROI (3-year)           81%             150%          200%

Even in the pessimistic scenario, NPV remains positive.

Common Finance Objections and How to Handle Them

Objection	Response
“Can you do it cheaper?”	“Here’s the minimum viable scope at €X. The trade-off is [what you lose].”
“Can it wait until next year?”	“Each month of delay costs €X in [lost revenue / increased incidents / team productivity].”
“How do you know these numbers?”	“Here’s the data source: [actuals from last 6 months / industry benchmark / vendor quote].”
“What’s the risk?”	“Here’s the sensitivity analysis. Even pessimistically, we break even in X months.”
“Can’t the existing team do this?”	“They can, but it means stopping [current project], which has [business impact].”

---

Anti-Patterns and Common Mistakes

1. The “Trust Me” Business Case

The mistake: Qualitative arguments without numbers. “We need to modernize because our architecture is old.” Why it fails: Finance allocates capital based on numbers, not feelings. “Old” isn’t a financial metric. Instead: “Our legacy architecture causes 4x more incidents than industry average, costing €180K/year in engineering time and €95K/year in lost revenue.”

2. Cherry-Picking the Metric

The mistake: Showing only ROI (which looks amazing) while hiding a 4-year payback period. Why it fails: CFOs see through this. It destroys your credibility for future asks. Instead: Show all four metrics. Let the strengths speak for themselves.

3. Ignoring Ongoing Costs

The mistake: Showing investment cost but not ongoing maintenance, support, and depreciation. Why it fails: Your “€200K one-time investment” actually costs €60K/year to maintain, which changes the NPV significantly. Instead: Model the full lifecycle cost: build + maintain + eventually decommission.

4. Confusing Savings with Revenue

The mistake: “This saves 10 engineer-hours per week” treated as €33K revenue. Why the CFO is skeptical: Those engineers’ salaries don’t decrease. The “savings” only create value if those hours are redirected to revenue-generating work. Instead: Be explicit: “10 hours/week freed up, redirected to Feature X, which drives €Y revenue.”

5. No “Do Nothing” Option

The mistake: Presenting only the investment option without analyzing the cost of inaction. Why it fails: Without a baseline, there’s no urgency. Instead: Always start with “Option 0: Do Nothing” and quantify its cost. This creates the urgency for action.

---

Real-World Examples

Example 1: AI Platform Investment (MMS-Relevant)

BUSINESS CASE — AI-Powered Service Desk

THE PROBLEM
Service desk handles 15,000 tickets/month. 40% are repetitive L1 queries
(password resets, order status, store hours). Each costs €8 to handle manually.
Total cost of repetitive tickets: €576,000/year.

THE PROPOSAL
Deploy an AI agent to handle the top 10 repetitive query types, targeting
50% automation rate within 6 months.

INVESTMENT:     €350,000 (3 engineers x 6 months + GCP AI costs)
ANNUAL SAVINGS: €288,000 (50% of €576K repetitive ticket cost)
PAYBACK:        14.6 months
3-YEAR NPV:     +€315,000 (at 10% discount rate)

Example 2: Tech Debt – Monolith Decomposition

BUSINESS CASE — Checkout Service Extraction

THE PROBLEM
Checkout is embedded in the monolith. Any change requires full deployment
(3 hours). We deploy checkout changes 2x/week = 312 engineer-hours/year
in deployment overhead alone. Plus 8 checkout-related incidents/year
(MTTR: 4 hours, impact: €15K/incident in lost sales).

INVESTMENT:     €280,000 (2 engineers x 5 months + infrastructure)
ANNUAL SAVINGS: €140,000 (deployment time) + €120,000 (reduced incidents)
PAYBACK:        13 months
3-YEAR NPV:     +€400,000

---

References

• Financial Intelligence for Entrepreneurs – Karen Berman & Joe Knight (2013) – NPV, ROI, and financial metrics for non-finance leaders
• The Art of Business Value – Mark Schwartz (2016) – Connecting IT investment to business outcomes
• Measure What Matters – John Doerr (2017) – OKR framework for linking investment to outcomes
• IBM Cost of a Data Breach Report – Annual report on breach costs (useful for security ROI)
• Gartner IT Score – IT investment benchmarking
• McKinsey Digital – Tech Debt – Research on tech debt quantification
• Wiring the Winning Organization – Gene Kim & Steven Spear (2023) – Connecting engineering improvement to business performance
• DORA Metrics – Quantifying engineering productivity for business cases