Define the decision first
ROI analysis should support a decision such as expanding a lifecycle program, replacing a tool, or improving production efficiency. The decision determines the period, costs, and outcomes to include.
Write the scope before collecting numbers so the analysis does not change to fit a preferred conclusion.
Use a transparent formula
A common structure is attributed return minus total email cost, divided by total email cost. State whether return means revenue, gross profit, contribution margin, or another measure.
Do not compare analyses that use different definitions without adjustment. Keep the source and calculation for every input.
Count the full cost
Include software, message volume, creative work, engineering, data operations, testing, compliance review, and allocated staff time where material. Setup and migration costs may need separate treatment.
Document shared-cost allocation instead of quietly excluding it. A range can be more honest than a false point estimate.
- Platform and sending fees
- Labor and agency work
- Data and integration costs
- Testing and quality assurance
- Governance and compliance operations
Choose an attribution approach
Campaign parameters, analytics events, and purchase records can connect email interactions with outcomes. The attribution window and model determine how much credit email receives.
Last-click, first-click, and multi-touch views answer different questions. Present the chosen method and compare alternatives when the decision is sensitive to attribution.
Look for incremental impact
Attributed revenue is not automatically caused by email; some recipients would have acted anyway. Where feasible, use an appropriate holdout or controlled rollout to estimate incremental impact.
Experiments must be designed carefully, with sufficient observation and ethical treatment. If a control is impractical, label causal uncertainty clearly.
Account for delayed and indirect value
Education, onboarding, retention, and service communication may influence outcomes beyond a short conversion window. Define those effects before analysis rather than adding them afterward.
Avoid assigning speculative monetary values to opens or clicks. Use observable downstream outcomes and show uncertainty.
Report scenarios, not false precision
Calculate conservative, central, and optimistic scenarios for uncertain attribution, margin, and cost assumptions. Sensitivity analysis shows which inputs actually change the decision.
Pair ROI with recipient-health measures such as complaints and opt-outs. A positive short-term result does not justify practices that undermine permission or long-term trust.
- Scope and period
- Return definition
- All material costs
- Attribution model
- Uncertainty range
- Recipient-health guardrails