How do I make a recommendation when the evidence is incomplete?
Quick answer
State the decision to be made, summarize the evidence you have, and name important gaps. Then give a recommendation that matches the strength of that evidence, explain your reasoning, and say what assumptions it depends on. Close with what information could change your view and, if useful, a review point. This lets readers act without mistaking a tentative judgment for a proven result.
The question, the details, the next step.
I need to recommend an option, but the information I have is incomplete. How do I avoid making it sound more certain than it is?
Make the uncertainty part of the recommendation rather than hiding it in a disclaimer. Separate three things: what the evidence directly shows, what you infer from it, and what remains unknown. Then make a recommendation whose strength fits those limits. You can still be useful and decisive: recommend a next step, while making clear that it is provisional and identifying what could change it.
A useful test is whether a reader can tell which statements are observations and which are your interpretation. Avoid turning a limited observation into a claim about a larger group, a future outcome, or a cause unless your information supports that claim.
How do I distinguish evidence from inference, and still recommend something before every gap is filled?
Evidence is the information you can point to in the material you actually have. Inference is your interpretation of what that information may mean. Unknowns are details that could matter but have not been established.
For example, in a hypothetical planning note, the evidence might be: “Two of the three people who replied preferred Option A.” An inference might be: “Option A may be more acceptable to the people who responded.” The replies do not establish what everyone prefers, so avoid writing “Everyone wants Option A.”
If you need to act now, connect your recommendation to its assumptions: “Based on the responses received so far, I recommend Option A for the next planning step, assuming the people who have not replied do not raise a conflicting requirement. This is provisional, not a finding about the full group.” That gives a direction without implying the missing information has been resolved. Consider how difficult the decision would be to reverse when explaining why this is an appropriate next step.
How can I explain what might change my recommendation without sounding as if I have no view?
Name specific decision conditions, not vague possibilities. Ask what new information would actually affect the choice. In the hypothetical example, you might write: “I would revisit Option A if the remaining responses identify a requirement it cannot meet.” That is clearer than “More information may change things.” You could also say what would strengthen your view, such as receiving no conflicting requirements by an agreed review point.
Focus on gaps that bear directly on the decision rather than listing every imaginable uncertainty. State the action clearly, then place the limits and change conditions beside it. This shows that you have a view while making its boundaries visible.
Can you give me a format I can reuse in a memo, and a quick way to check the wording?
Use this structure, filling each bracket only with information you can support:
Decision: [What needs to be decided, and by when?] Evidence available: [What has been observed, received, or confirmed?] Limits: [What is missing or uncertain?] Recommendation: [What do you recommend now?] Reasoning: [How does the evidence support this choice, and what assumptions are you making?] Change conditions: [What specific new information would make you reconsider?] Next step or review point: [What should happen next, and when should this be checked again?]
For the example, the recommendation could be: “Proceed with Option A for the next planning step.” The limits and change-condition lines clarify that the replies are incomplete and that a conflicting requirement would prompt a review. Keep those details hypothetical unless they match your actual evidence.
Before sending, ask of each sentence, “What supports this?” Use phrases such as “so far” or “based on the information available” when they accurately describe the evidence. Do not bury a clear recommendation under repeated hedges. Check that the reader can identify what you recommend, why, what you do not know, and what would trigger a change.
A practical checklist
Name the decision and the time or situation it applies to.
Separate observed evidence, interpretation, and unresolved gaps.
State the recommendation directly and label it provisional when appropriate.
Explain the assumptions connecting the evidence to the recommendation.
Name the specific information or condition that would prompt reconsideration.
Set a next step or review point if the decision needs follow-up.
A few more useful answers
Should I wait for complete evidence before making any recommendation?
Not necessarily. If a decision is needed now, give a recommendation proportionate to what is known, identify important gaps, and state what would prompt a review. Whether to proceed depends on the decision and its consequences.
How many uncertainties should I mention?
Include the gaps that could materially affect the recommendation. A short, relevant explanation is usually clearer than a long inventory of unknowns that would not change the decision.
Can a recommendation be confident and still provisional?
Yes. You can be clear about the action you recommend now while being candid that the evidence is incomplete. Confidence in a next step is not the same as claiming the outcome is proven.