Sizing, ROI & Sample Calculators Pack

Quick overview

This pack helps teams rapidly judge which ideas deserve attention, how big an opportunity might be, how quickly an investment could pay back, and how large an experiment needs to be to learn reliably. Each calculator includes the core formula, a worked example, recommended assumptions to validate, and practical sensitivity checks so you avoid overconfidence from single-point estimates.

What's included

• Opportunity sizing worksheet (frequency × cost/time per event)
• Simple ROI calculator (annualized benefit vs. investment + payback)
• Sample-size calculator for common test types (proportions and means) with explanation and examples

1) Opportunity sizing worksheet

Purpose: convert a qualitative idea into a rough annual value so you can prioritize. Use conservative assumptions and then run sensitivity checks.

Core formula:

Annual Opportunity = Frequency per period × Periods per year × Benefit (cost, time, or revenue) per event

Worked example:

• Issue: Reduce handling time for an order. Estimated time saved per order = 3 minutes (0.05 hours).
• Orders per day (site): 400 → Days per year: 250 → Frequency per year = 400 × 250 = 100,000 orders.
• Labor cost per hour = $25 → Value per order = 0.05 × $25 = $1.25.
• Annual opportunity ≈ 100,000 × $1.25 = $125,000.

Recommended assumptions to record and validate:

• How you measured frequency (sample period, system logs, manager estimate)
• How you estimated per-event benefit (time measured, scrap avoided, price uplift)
• Whether the benefit is recurring or one-time

Sensitivity check: recompute with low/medium/high scenarios (e.g., time saved = 1, 3, 5 minutes) and note how priority changes.

2) Simple ROI & payback calculator

Purpose: quickly judge whether an investment in a pilot or solution is justified.

Core numbers to collect:

• Annual benefit (from opportunity sizing) – B
• One-time implementation cost – C (or first-year cost if recurring)
• Annual ongoing cost – O (if applicable)

Key metrics:

• Simple ROI (%) = (Annual net benefit / First-year cost) × 100, where Annual net benefit = B − O and First-year cost = C + O (if costs are split)
• Payback period (years) = First-year cost / Annual net benefit (useful when benefit and cost timing are simple)

Worked example:

• Annual benefit B = $125,000
• One-time pilot cost C = $40,000, ongoing yearly cost O = $5,000
• Annual net benefit = 125,000 − 5,000 = 120,000
• First-year cost = 40,000 + 5,000 = 45,000
• Simple ROI = (120,000 / 45,000) × 100 ≈ 267% ; Payback ≈ 45,000 / 120,000 ≈ 0.38 years (~5 months)

Notes & cautions:

• If benefits or costs are spread, consider using multi-year cashflows and a discount rate for NPV/IRR analysis.
• Record major risks that could reduce benefit (adoption, operational impact, regulatory constraint).
• Run conservative and pessimistic scenarios: assume only 50% of estimated benefit for a conservative ROI.

3) Sample-size calculator for experiments

Purpose: estimate how many users, events, or observations you need per variation to detect a plausible effect with acceptable confidence and power.

Core concepts to capture: baseline rate (p), minimum detectable effect (d, absolute), significance level (α, commonly 0.05 → Z≈1.96), statistical power (1−β, commonly 0.8 → Z≈0.84).

Approximate formula for proportions (A/B test comparing conversion rates):

n per group ≈ 2 × (Z_{1−α/2} + Z_{1−β})^2 × p × (1 − p) / d^2

Where:

• p = pooled estimate of baseline conversion (use baseline if unsure)
• d = absolute difference you want to detect (e.g., 0.01 for +1 percentage point)
• Z_{1−α/2} = 1.96 for 95% confidence; Z_{1−β} = 0.84 for 80% power

Worked example (detecting a lift on a 5% baseline):

• Baseline p = 0.05, target lift = +20% relative = +1 percentage point → d = 0.01
• n ≈ 2 × (1.96 + 0.84)^2 × 0.05 × 0.95 / 0.01^2
• Calculate (1.96+0.84)^2 ≈ 7.84; so n ≈ 2 × 7.84 × 0.0475 / 0.0001 ≈ 2 × 7.84 × 475 ≈ 7,448 per group → total ≈ 14,896
• Interpretation: small absolute effects on small baselines need large samples.

Rule of thumb and alternatives:

• When you lack a reliable baseline, use p = 0.5 for a conservative (largest) sample-size estimate.
• For continuous metrics (means), a comparable formula uses the standard deviation σ and the absolute difference d: n per group ≈ 2 × (Z_{1−α/2} + Z_{1−β})^2 × σ^2 / d^2
• Practical minimum: if your expected total sample per variation is under a few hundred, small percentage-point effects will be undetectable; consider larger effect thresholds or different experiment designs.

Experiment hygiene:

• Avoid early peeking and optional stopping unless using proper sequential methods (which require different sample calculations).
• Pre-register your metric, baseline, minimum detectable effect, α, and power before running the test.
• If tracking sessions/events rather than users, ensure independence or adjust calculations for clustering.

Practical checklist before you run a test or invest

1. Document key assumptions (frequency, baseline, unit value, costs).
2. Run low/medium/high sensitivity scenarios and check whether priority changes.
3. Compute sample size for the realistic minimum detectable effect; if required sample is unreachable, consider raising the MDE or running a pilot for variance estimates.
4. Estimate adoption risk and operational constraints that could reduce realized benefit.
5. Record decision criteria: what ROI or payback would justify scaling?

Spreadsheet recipe (copy into Excel or Google Sheets)

Setup recommended columns and cells for the Opportunity + ROI worksheet:

• Cell A1: "Frequency per period"; A2: numeric input
• Cell B1: "Periods per year"; B2: numeric input
• Cell C1: "Value per event"; C2: numeric input
• Cell D1: "Annual opportunity"; D2 formula: =A2*B2*C2
• Cells for One-time cost, Annual ongoing cost, Annual net benefit, ROI%, Payback with simple formulas shown earlier.

For sample size, implement the proportions formula using spreadsheet functions for Z values (1.96 and 0.84 by default) and allow inputs for p and d so you can easily toggle scenarios.

How this supports discovery & innovation

These calculators convert ideas into defensible numbers quickly so teams can prioritize experiments that are affordable, measurable, and likely to matter. They also help avoid common traps—over-optimistic single-point estimates, underpowered tests, and investments without clear adoption risks captured.

Next steps & capability suggestions

Copy this worksheet into your site or team domain. Consider adding an interactive version of these calculators that computes results in real time, saves scenarios, and stores submissions so you can compare proposals over time.