In this guide
- 01Name the decision
- 02Compare credible alternatives
- 03Expose evidence and uncertainty
- 04Record the commitment
A decision memo helps an authorized reader choose among credible alternatives. State the decision, your recommendation, the evidence, the main trade-off and the commitment being requested. Put supporting detail where the reader can inspect it without hiding the choice in a long narrative.
Use a memo when a decision depends on reasoning that should be reviewable before and after the meeting. It complements a weighted decision matrix, which makes some comparison judgements visible, and a premortem, which examines how a plan could fail.
Identify the decision and the person who can make it
Write the question as an action within someone’s authority. “What should we do about packaging?” is broad. “Should the operations director authorize a bounded returnable-container trial for one defined delivery route?” identifies the proposed commitment.
NASA’s decision-analysis guidance separates the decision context, alternatives, criteria, evaluation and recommendation. It is an engineering reference, but that structure is useful for a business memo. The document should explain the actual choice rather than merely describe the project’s history.
Check the decision maker’s authority. A reader may be able to release staff time but not change a supplier contract. If the proposal depends on both, make the dependency explicit and identify the further decision. An apparent approval is weak if the requested commitment cannot be delivered.
Build the memo around one decision
Consider a hypothetical parcel service deciding how to investigate reusable containers. The team has two physical concepts, some exploratory observations and important unanswered questions about returns. A full rollout would exceed the available evidence; a limited trial might be reasonable.

Start with a short recommendation and its conditions. Then describe the alternatives fairly. Include continuing the current arrangement where it is a real option, together with its implications. Do not compare a fully developed preferred proposal with an implausible or deliberately weak alternative.
A useful original memo structure is:
| Part | What the reader needs |
|---|---|
| Decision requested | The action, authority and timing |
| Recommendation | The preferred option and why it is proportionate |
| Alternatives | Credible choices, including a smaller commitment where useful |
| Evidence | Observations and sources supporting the comparison |
| Trade-off | What is given up or made harder |
| Uncertainty | What remains unresolved and how it affects the choice |
| Commitment | People, time, spending and operating responsibility |
| Review condition | What will trigger continuation, revision or stopping |
This structure is a writing aid, not a formal government business-case template. A narrow decision may fit a page; a consequential proposal may need appendices and specialist input. Brevity should reveal the reasoning rather than omit a material condition.
Separate what is known, estimated and preferred
Label observations as observations and estimates as estimates. A prototype test may show how a container was handled under specified conditions. It does not establish annual savings, future demand or the reliability of a return network. Explain which further assumptions are needed to reach those conclusions.
Use a simple evidence statement: what was examined, in which setting, with what result and limitation? Link directly to the relevant source or internal record the authorized reader may access. Avoid a long reference list that leaves the connection to the recommendation unexplained.
HM Treasury’s MCDA guidance emphasizes transparent criteria and stakeholder consideration in options appraisal. When the fictional team compares container options, distinguish required conditions from preferences and uncertain performance. If changing a plausible weight reverses the recommendation, show that sensitivity instead of presenting the total as decisive.
Keep value judgements visible. Choosing service continuity over a possible saving is not necessarily a factual error. The decision maker needs to understand the trade-off and decide whether it fits the organization’s objectives. A calculated score cannot make that responsibility disappear.
Ask for a bounded commitment
The fictional memo might recommend examining one concept on a defined route, with a named operating owner and a review before expansion. Specify what the trial would teach and what it would leave unresolved. Do not promise a rollout decision if the planned evidence cannot support one.

An illustrative recommendation could read: authorize a supported trial of the stackable concept, provided the return-handling responsibility and the participant arrangements are agreed before it starts. Review handling effort, successful returns and operational exceptions before deciding on another route. This is a fictional proposed commitment, not a reported approval or result.
Name the resources that matter. Spending is only one part of the commitment. A trial may need a technician, a receiving area, support staff and attention from a manager who already has a full workload. Innovation budgeting helps connect those resources to the next question.
State what would cause reconsideration. A missing operating owner may block the trial before it begins. An observed failure may trigger a pause. Evidence that the concept can be handled may justify another test while leaving economics open. These conditions should be understandable before the team sees the results.
Explain how results will be interpreted
The Magenta Book’s evaluation guidance distinguishes examining implementation and outcomes from making a causal claim. Use that distinction when a memo requests a pilot. A pilot can reveal operational problems without proving that the proposal caused a measured improvement.
If the team needs causal evidence, describe the comparison and the limitations or seek appropriate evaluation support. A favourable before-and-after chart is insufficient when other relevant conditions changed. The evidence-checking guide gives questions for examining such claims.
Do not make uncertainty unreadable through a long risk register. Identify the few uncertainties that could materially change this decision, state what is being done about them and link to the detailed record where needed. A reader should be able to distinguish an unresolved assumption from a managed operating condition.
Write for the reader’s task
GOV.UK’s writing guidance emphasizes clear language and structure that helps people find and understand what they need. Applied to a memo, lead with the requested decision and use descriptive headings. Explain an unfamiliar term when it first matters.
Remove background that does not change the choice. Preserve context that explains a material constraint or a credible alternative. Give figures a unit, a time period and a source. A precise number without those details can make the proposal look more certain than the evidence warrants.
Ask a colleague unfamiliar with the project to identify the decision, recommendation and strongest objection. Their difficulty is useful editing evidence. It does not prove the recommendation is wrong, but it shows where the memo is failing to communicate the reasoning.
Review limits and preserve the actual decision
This guide combines methodological and institutional sources with an original hypothetical example. It does not establish that a particular page count improves decisions or replace the organization’s formal approval requirements. The proposed trial has not been performed.
After the decision, record what was actually authorized, by whom, with which conditions and when it returns for review. Keep a material disagreement visible. Use the memo later to compare new information with the assumptions accepted at the time, rather than rewriting the history around the eventual outcome.
Sources:
- NASA (n.d.). 6.8 Decision Analysis.
- HM Treasury (2024). Use of Multi-Criteria Decision Analysis in options appraisal of economic cases.
- HM Treasury (2026). Magenta Book: Central Government guidance on evaluation.
- Government Digital Service (n.d.). Writing guidelines.



