A decision memo for the board: fit the reasoning, alternatives and risk into a short document
Write a board decision memo with a recommendation, real alternatives, evidence, risk and the scope of approval. Turn market research into a concrete choice about further work.
Article language: English

In this article
The board opens a report, and the report opens another sixty pages in front of them. Everything is described, yet the decision still waits for a paragraph of its own. A short decision memo is meant to solve that problem: show what you propose, why the choice makes sense, what the alternatives are and what remains unknown. Appendices can stay extensive. The route to the point should be short.
A good memo is not a summary of every activity completed. It organises the material around a choice for which someone will be responsible. It also allows the reader to reject a recommendation for the right reason. If the only options are to sign the request or repeat the whole research independently, the document has not done its job.
Open with the choice you need
The first paragraph should state the decision question, recommendation and scope of approval. Take a fictional Polish company considering entry into Japan with a technical service. The recommendation may concern conversations with one type of partner, not a full sales launch. Those scopes must not be interchanged, even if both look good on a slide.
An example: “We recommend preparing and conducting a limited series of conversations with distributors serving the specified segment. We want to establish whether the support they require is within the company's capabilities. Approval covers the discovery stage and its budget; the sales decision will return after the results have been assessed.” Specific limits must be filled in before approval.
This opening immediately lets the recipient judge whether they are the right person to decide. If the document only needs to report research results, use the format described for closing a market study and recommending next steps to the board (in Polish). A decision memo goes further: it explicitly asks for a choice.
Separate the situation from the story of your own work
Context explains why a decision is needed now. It does not have to begin with the date of the first conversation with the research provider or the history of successive presentation versions. The board needs to understand the opportunity or difficulty facing the company and what will happen if the decision is postponed.
In the Japan example, the context might be the need to establish a support model before preparing an offer. Interest in the market alone does not yet justify investment. State whether the basis is current enquiries, an identified segment, available resources or a strategic plan. The source material should lead to the market research brief (in Polish), so the reader understands the limits of the original question.
Use three separate sentences: what we know, what follows from it and what we propose to check. For example: the company profiles analysed describe a certain service scope; their fit with our service remains unconfirmed; therefore conversations about a specific application are needed. A mention in a directory does not quietly become evidence of demand.
Present alternatives that can be chosen without embarrassment
A weak memo offers one reasonable proposal and two decorations: an extremely expensive option and complete inaction. A good one presents real paths with different consequences. The reader should be able to choose another option without feeling they have just voted against common sense.
For the fictional project, compare:
- conversations with a selected group of partners to test the cooperation model;
- refining the offer and materials before contact, at the cost of learning market reactions later;
- postponing Japan and using resources in another segment already served;
- abandoning the application under consideration while retaining the possibility of exploring another need.
Assess every option against the same criteria: expected information or benefit, effort, constraints and the ability to withdraw. Choosing between a partner model and direct sales requires a different comparison from choosing a catalogue language. The separate considerations on distributors and direct sales (in Polish) can help.
If “wait” has a cost, name it honestly. It might mean continuing to maintain makeshift materials or losing team time to inconsistent answers. If the cost has not been estimated, say so. Do not add a dramatic lost opportunity merely to make the recommendation look more attractive.
Support the recommendation with evidence of the right strength
For each key argument, provide the evidence and its limit. A source may confirm the scope of a company's activities, but not its willingness to talk. An interview may explain one participant's problem, but not the size of an entire market. A number is meaningful only with its definition, period and scope of observation.
Instead of saying “the market is promising”, show the reasoning: which segment was analysed, why it was selected for further checking and what could reverse the recommendation. When results conflict, use the distinctions in the article on conflicting sources in Asian market research (in Polish). Removing an inconvenient number does not resolve a difference in definitions.
If a competitor comparison supports an argument about a gap in the offer, connect the conclusion to the Korean competitor analysis (in Polish) as an example of comparing scope and evidence. A feature missing from a competitor's website does not mean nobody offers it. The memo should show that limitation too when the choice depends on it.
There is no need to move every quotation into the main text. A source appendix can hold details, while the memo points to specific findings. “Documentation in the folder” is too broad a pointer. The reader should not have to play hide-and-seek with the argument they are being asked to accept.
Borrow the structure, not another organisation's obligation
GDS Way describes recording architectural decisions through context, decision, status and consequences, preserving the reasoning even when team membership changes. This is a technical ADR, or architecture decision record. A business memo can use the separation of reasons from consequences without copying a software team's tools. The page indicates that its own review date has passed, so we use it as a description of a method without assuming every detail of the GDS process is current.

In practice, mark the status clearly: “recommendation for consideration”, “accepted” or “superseded by a later decision”. The word “decision” in a filename does not prove anyone has made one. A PDF does not acquire managerial authority when it is exported.
The UK government's architectural decision record framework distinguishes decision levels by the reach of their effects and the need to escalate when a choice concerns wider dependencies. The document is intended for the UK public sector. For a company it offers a useful control question: does the person approving the proposal have authority covering all its consequences?
Establish that before circulating the memo. A person leading expansion might approve the scope of conversations, while a change to the service team's workload could require another decision. The request should identify both dependencies rather than assuming one signature automatically settles the rest.
Describe risk so someone can work with it
“Market risk” fits almost anything and therefore helps little. Describe the event, its effect and how to reduce it. In our example, if a partner expects support during hours the company cannot cover, the cooperation proposal needs to change. The conversations are intended to establish requirements before making a promise about support.
Do not equate reducing a risk with removing it. Record separately what the company can check in the next stage and what will remain uncertain until work actually starts. Verifying a potential partner (in Polish) helps when assessing contacts, but even careful source checks cannot guarantee a successful future relationship.
Add a condition for returning to the decision: a cost change, a missing critical skill or a test result that challenges an assumption. The project stopping criteria (in Polish) can help. Conditions set before starting make it easier to assess new knowledge without adjusting the rules to a preferred option.
Walk the reader through one recommendation
Continue the fictional case of entering Japan. The company can perform the service but does not know whether a future partner will expect broader local support. The memo can recommend conversations solely to establish requirements and compare them with the company's capacity. It should not promise at this stage that the chosen path will lead to a contract. The next stage produces information for a decision, not guaranteed sales.
In the resources section, the author identifies preparation of material, language work, a technical person's participation and analysis of findings. In the boundaries section, they state that conversations do not include agreeing final commercial terms. In the results section, they list the questions to answer: expected support, the people involved in the assessment and whether further testing of the application is feasible. This allows the effort to be compared with the information the company wants to obtain.
If the recommendation requires a change to a way of working established by headquarters, show what follows from local data and what is a company-wide rule. The relationship between headquarters and a local branch (in Polish) helps distinguish a decision about a particular application from an unnecessary argument about the organisation's entire standard.
Alongside the recommendation is a fair alternative: refine the offer without conversations, using current research. This costs less to coordinate but leaves uncertainty about expected support. A second alternative is to postpone the project and use the team's skills for existing clients. That may be reasonable if the team is unavailable. The author should show why, despite that, they recommend limited contact with partners.
The final request leaves room for a decision: accept the proposal, choose an alternative or return the document with a specific question. The responsible person and date appear beside it. If the board changes the scope, record the change at the request itself, not only in a message comment. The people doing the work must distinguish the approved version from the original request, especially when both later circulate as attachments.
Test the recommendation with an uncomfortable question
Before sending, choose the argument on which the recommendation depends most heavily and temporarily remove it. Does the choice still make sense? If so, state which other reasons support it. If not, name the dependency in the main text. This authorial exercise helps reveal a conclusion that appears to rest on several reasons but actually stands on one unconfirmed assumption. The reader should see that structure without dismantling the report themselves.
Next, ask another person to formulate the strongest argument for an option not selected. This is not a game against the author. It tests whether the document can fairly present the cost of its own recommendation. If the argument is strong, add it and explain why it does not prevail under current conditions. The reasoning then remains useful later, when conditions change and the company considers another direction.
Shorten the route to assessment, not the meaning
There is no single correct page count for every choice. Start with a short main section and move details into organised appendices. If omitting a condition could reverse the decision, that condition belongs in the main section. Saving space does not justify hiding the most important limitation in a footnote.
Ask someone outside the project to read the document as a test. Without further explanation, have them identify the approval requested, an alternative, what has not been confirmed and what happens after approval. If their answers differ from the author's, revise the structure or sentences. A longer oral presentation should not be mandatory operating instructions for the memo. If the comments concern the specialist basis of the recommendation, describe the specific problem and question for the expert (in Polish) instead of replacing an assessment of the reasoning with personal preference.
Take a list of disputed assumptions to the decision meeting (in Polish) rather than planning to read the whole document aloud. After the choice, record the result in the project decision log (in Polish), preserving the earlier recommendation and the reason for any change. A reasonable change of position is part of the work, not a stain on the author's honour.
If you need to turn scattered data about Korea or Japan into material for choosing the next stage, see the Aeneia Oriens offer. Prepare the board's question, current sources and one decision that has stalled; then the work can be matched to the real unknown rather than the thickness of a future report.







