ROI Modelling

Purpose

Project the financial return of proceeding from A3 to full A4–A7 development, comparing expected revenue and cost savings against investment required—producing Net Present Value (NPV), Internal Rate of Return (IRR), and payback period estimates that ground the viability lens in financial evidence rather than optimism.

ROI modelling in A3.4 differs from A1.4 business case modelling in one critical way: A1.4 projected from assumptions; A3.4 projects from evidence. Pilot data (actual retention, actual support costs, actual conversion rates) replaces the assumptions that drove the original business case, producing higher-fidelity financial projections.

When to Use

Use ROI modelling when:

• Conducting A3.4 viability assessment (mandatory)
• A3.5 governance forum requires financial justification for investment
• Comparing multiple prototypes on financial merit
• Building A4 budget requests with evidence-based projections

Do NOT use when:

• Insufficient data—if pilot had <25 users or <4 weeks, financial projections are fiction dressed as spreadsheets
• Strategic investments where ROI is not the primary lens (e.g. defensive innovation, regulatory compliance)—acknowledge strategic rationale separately

Sample Size and Duration

Effort: 12–20 hours per prototype (Financial Analyst: 8–12 hours; Product Owner: 4–8 hours)

Duration: 2–3 days within A3.4 timebox

Prerequisites: A3.3 financial data available (actual costs, usage-derived revenue potential)

Prerequisites

• A3.3 pilot data: Actual retention rates, usage patterns, conversion rates (if monetised), support costs, infrastructure costs
• A1.4 market data: Total addressable market (TAM), serviceable addressable market (SAM), serviceable obtainable market (SOM)
• A4–A7 cost estimates: Development cost, launch cost, operational cost (from Solution Architect and Finance)
• Revenue assumptions: Pricing model, conversion rates, expansion revenue
• Discount rate: Organisation's cost of capital or hurdle rate (from Finance)
• Time horizon: Typically 3–5 years for innovation projects

Complete Procedure

Step~1: Build Unit Economics (2–4 hours)

Start bottom-up with per-customer economics:

where ARPU is Average Revenue Per User (monthly) and Churn Rate is monthly churn.

Example (C001 Smart Checkout):

• A3.3 pilot: 62% Week~8 retention estimated monthly churn ≈ 8%
• Projected ARPU: $15/month (from pricing research + pilot willingness-to-pay signals)
• LTV = 15 × (1 / 0.08) = 187.50
• Estimated CAC: $45 (from comparable SaaS benchmarks, adjusted for pilot acquisition data)
• LTV/CAC = 4.2:1 (green!60!blackPass: ≥3:1)
• CAC Payback = 45 / 15 = 3 months (green!60!blackPass: ≤12 months)

Step~2: Project Revenue (2–4 hours)

Build monthly revenue model for 36-month horizon:

• Month~1–6 (A4 development): $0 revenue, full development cost
• Month~7–9 (A7 launch): User acquisition ramp. 100 500 2,000 users.
• Month~10–36 (growth): Apply growth rate (conservative, expected, optimistic scenarios)

Step~3: Project Costs (2–4 hours)

Step~4: Calculate NPV (1–2 hours)

where C_t is net cash flow in period t, r is the discount rate, and T is the time horizon.

C001 example (expected case, r = 12%, 36-month horizon):

• Total investment (A4–A7): $1.16M
• Projected 36-month revenue: $2.7M
• Projected 36-month operating costs: $2.0M
• NPV: +$1.4M (after discounting)
• IRR: 38% (exceeds 12% hurdle)
• Payback: Month~18 (within 24-month target)

Step~5: Run Three Scenarios (2–4 hours)

Decision interpretation:

• Expected and best cases are strong
• Worst case: NPV -200K, IRR below hurdle —but loss is limited (200K, not $2M)
• Risk-adjusted verdict: Positive expected value with bounded downside

Feed scenarios into sensitivity analysis (the referenced method) to identify which assumptions matter most.

Step~6: Document Assumptions and Limitations (1–2 hours)

Every ROI model is only as good as its assumptions. Document explicitly:

• Which inputs are pilot-validated (retention, support cost) vs. assumed (growth rate, CAC at scale)?
• What market conditions are assumed? (competitor actions, regulatory changes, economic climate)
• What is the error range? Cooper notes A3-stage estimates carry 50–100% error —present ranges, not point estimates

Quality Criteria

1. Pilot-grounded: Key inputs (retention, support cost, usage) derived from A3.3 data, not assumptions
2. Three scenarios: Worst, expected, best cases presented—never a single-point projection
3. Assumptions explicit: Every input documented with source (pilot data, benchmark, assumption)
4. Error acknowledged: 50–100% error range at A3 stage stated clearly
5. Unit economics first: LTV/CAC viable before running full NPV model
6. Time-discounted: NPV uses appropriate discount rate (not undiscounted totals)
7. Reviewed by Finance: Financial analyst validates model structure and assumptions

Theoretical Foundation

Seminal references:

• : The standard corporate finance text establishing NPV as the primary investment evaluation method. NPV discounts future cash flows to present value using the cost of capital, answering: “Is this project worth more than it costs?” A positive NPV means the project creates value; negative means it destroys value.
• : Applied financial analysis to innovation investment decisions within the stage-gate framework. Key insight: financial projections at early gates are unreliable (error 300%); by A3.4, pilot evidence reduces error to 50–100%—still uncertain, but actionable with scenarios.

Contemporary references:

• : Positioned unit economics (CAC, LTV, payback period) as the viability metrics startups and innovation teams should model first—before complex DCF models. If unit economics don't work, NPV is irrelevant.
• : Connected ROI modelling to business model structure: revenue streams, cost structure, customer segments, and channels all feed the financial model. A3.4 ROI must reflect the validated business model, not the assumed one.

Key Financial Metrics

Challenges and Solutions

Challenge 1: Optimism Bias in Projections

Symptoms: Growth rate assumes 30%/month sustained for 3 years. Team anchors on best-case scenario. CFO calls the model “fiction.”

Solutions: Ground growth assumptions in pilot data and market comparables. If pilot acquired 100 users in 6 weeks, 30%/month sustained growth is aspirational, not expected. Present worst case first; if worst case is acceptable, best case is upside.

Challenge 2: Pilot Economics Production Economics

Symptoms: “Support costs were only 5K for 100 users—so 50K for 1,000 users!” Support costs don't scale linearly; neither does infrastructure.

Solutions: Model scaling non-linearity explicitly. Support: 1:50 ratio in pilot 1:200 in production (requires self-serve investment). Infrastructure: economies of scale reduce per-user cost but total cost increases. Use step-function cost models, not linear extrapolation.

Challenge 3: False Precision

Symptoms: “NPV is $1,423,847.” Implies precision that doesn't exist. Decision-makers treat it as exact.

Solutions: Round to significant figures (1.4M, not 1,423,847). Always present as range (NPV: 1.0M–1.8M expected, with -200K–4.2M full range). Use sensitivity analysis to show which inputs the NPV depends on most.

Relationship to Other Methods

ROI Modelling receives input from:

• Three-Lens Evaluation (the referenced method): Viability lens structures the financial assessment
• Cohort Analysis (the referenced method): Retention data drives LTV calculation
• Customer Success Management (the referenced method): Actual support costs feed operating expense projections

ROI Modelling provides input to:

• Sensitivity Analysis (the referenced method): ROI model becomes the base model for sensitivity testing
• Risk Assessment Matrix (the referenced method): Financial risks identified during modelling (e.g. worst-case NPV negative)
• Governance Forum (the referenced method): NPV, IRR, payback are primary financial inputs to A3.5 decision
• A4 Charter: ROI model becomes the financial baseline for A4 budget and success criteria

Tools and Templates

• Modelling: Excel, Google Sheets (standard for financial models); Causal, Mosaic (collaborative financial planning)
• Templates: SaaS financial model templates (Christoph Janz, Standard Metrics); custom A3.4 ROI template
• Scenario analysis: Excel Data Tables, Monte Carlo simulation (Crystal Ball, @RISK for advanced analysis)
• Benchmarks: SaaS Capital benchmarks, OpenView metrics database, Bessemer BVP Nasdaq index

• A. Croll & B. Yoskovitz (2013). Lean Analytics: Use Data to Build a Better Startup Faster. O'Reilly Media.
• A. Osterwalder & Y. Pigneur (2010). Business Model Generation: A Handbook for Visionaries, Game Changers, and Challengers. John Wiley & Sons.
• D. Kahneman (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
• R. A. Brealey, S. C. Myers & F. Allen (2020). Principles of Corporate Finance. 13 ed. McGraw-Hill Education.
• R. G. Cooper (2017). Winning at New Products: Creating Value Through Innovation. 4 ed. Basic Books.