Write an executive summary that supports a decision

An executive summary should let its intended reader understand the problem, the main findings and the action proposed without reconstructing the whole report. It is usually easier to write once those elements are settled in the report itself.

That does not make the summary a shortened introduction. An introduction tells the reader what the report will do. An executive summary reports what the completed work did, found, concluded and recommended. If the report does not establish a claim, the summary cannot safely add it.

If you are still deciding which form the assignment requires, first compare the structural jobs in our essay-or-report guide.

Start with the assignment instructions. Some reports require an executive summary; others use an abstract, a synopsis or no separate summary. The expected content and length are controlled by the brief, discipline and reader, not by one universal percentage.

Identify the decision before selecting details

Write one private sentence that names what the reader must understand or decide after reading the summary:

The centre manager needs to decide whether a limited reminder pilot is justified by the completed report.

This sentence is a selection rule, not necessarily a line for the final submission. Details belong in the summary when they materially affect that decision. Background that merely makes the report sound comprehensive can stay in the body.

RMIT University’s annotated report example presents the executive summary in the same broad order as the report: aim, method, main result, conclusion and recommendation. Its business-report guidance also describes the summary as an overview of the whole report that is written last. Australian National University’s report guidance similarly identifies context and aim, main findings, the key message and recommendations as typical components.

Treat those components as questions:

What problem, purpose or decision framed the report?
What evidence or approach produced the findings?
Which findings most affect the decision?
What conclusion follows from those findings?
What action is recommended, and what limit changes its strength or scope?

Work from a completed fictional report

The example below is invented teaching material. Meadowbridge Community Centre, its records, respondents and results are not real, and the numbers are not evidence about reminder systems elsewhere.

The fictional report asks whether Meadowbridge should pilot optional text-message reminders for free skills-workshop bookings. Its completed sections contain this information:

Purpose: assess whether a limited reminder pilot is justified by unused booked places.
Approach: review the centre’s fictional 12-week booking log and staff cancellation notes, then examine a voluntary feedback form completed by 40 workshop users.
Finding 1: 48 of 240 booked places were unused, or 20%.
Finding 2: 29 of those 48 unused-place records had no cancellation recorded.
Finding 3: seven of the 40 feedback respondents said they had forgotten at least one workshop date; six said they would not prefer text reminders.
Limits: the voluntary form does not represent every workshop user, the records do not establish why each unused place occurred, and the report did not test whether reminders change attendance.
Conclusion: unused booked capacity exists, but the evidence does not show that forgetting caused most non-attendance or that text reminders will solve it.
Recommendation: run an eight-week, opt-in pilot alongside the existing email reminder; compare non-attendance and cancellation patterns with the 12-week baseline and review accessibility, consent, staff time and participant feedback before any wider decision.

The arithmetic is deliberately inspectable: 48 divided by 240 is 0.20, or 20%. The example does not convert seven voluntary responses into a claim about all users, and it does not predict a reduction the report has never measured.

Draft the summary from the evidence map

Here is a possible executive summary for that finished report:

This report assessed whether Meadowbridge Community Centre should pilot optional text-message reminders for free skills-workshop bookings. It reviewed a fictional 12-week booking log and staff cancellation notes and examined a voluntary feedback form completed by 40 workshop users. Forty-eight of 240 booked places were unused (20%), and 29 of those unused-place records had no cancellation recorded. Seven feedback respondents reported forgetting at least one workshop date, while six said they would not prefer text reminders. These records show unused capacity but do not establish the cause of most non-attendance or whether reminders would improve it; the voluntary feedback also cannot represent every user. The report therefore recommends an eight-week, opt-in text-reminder pilot alongside existing email, followed by a comparison of non-attendance and cancellation patterns and a review of accessibility, consent, staff time and participant feedback before any wider decision.

The recommendation is direct, but its confidence matches the evidence. It proposes a reversible test rather than presenting an untested intervention as a proven solution.

Trace every summary sentence back to the report

Use a trace table after drafting. If a sentence has no home in the report, either add the necessary evidence to the report and analyse it there or remove the sentence from the summary.

On a narrow screen, scroll the table sideways to read every column.

Where each sentence in the fictional executive summary comes from
Summary sentence	Report source	Audit question
Purpose and decision	Introduction: commissioned purpose and decision	Does it name the actual report task rather than a broader topic?
Approach and evidence base	Methods/evidence-base section	Are the period, records and respondent group stated accurately?
Unused-place result	Findings: booking-log count	Do the numerator, denominator and percentage agree?
Feedback result	Findings: voluntary feedback	Does the wording identify respondents rather than generalise to all users?
Conclusion and limits	Discussion, limitations and conclusion	Does it separate what the records show from what they cannot establish?
Pilot recommendation	Recommendations	Is every action supported, scoped and capable of evaluation?

This audit also exposes a common problem: a polished sentence may still be unsupported. “Text reminders will substantially reduce missed bookings” sounds decisive, but the fictional report contains no intervention test and no effect estimate. The accurate version is the limited pilot recommendation above.

Keep findings, conclusions and recommendations distinct

These three moves are connected but not interchangeable:

A finding reports relevant evidence: 48 of 240 booked places were unused.
A conclusion answers what that evidence means for the report purpose: there is unused capacity, but its causes are not established.
A recommendation proposes a justified next action: test an opt-in reminder for a defined period and evaluate specified outcomes.

Jumping from a finding to a broad recommendation usually hides an inference. Ask what else would need to be true for the action to work. In this example, text access, consent, delivery reliability, participant preference and staff capacity matter. The report need not resolve every uncertainty, but the summary should not conceal the uncertainties that affect the decision.

If the report has not yet justified its preferred response, use the option-to-action recommendation guide to compare feasible alternatives before condensing the decision here. The summary reports that reasoning; it does not replace it.

Put important information first

A busy reader should not have to wait until the final line to learn the result. A useful order is often:

purpose or decision;
evidence base;
decisive findings;
conclusion with material limits; and
recommendation or next action.

Change that order if the brief or professional convention requires another one. For example, a reader-facing organisational report may lead with the recommendation and then supply its evidence. Whatever the order, keep the chain visible: evidence supports conclusion; conclusion supports action.

Prefer exact nouns and verbs over promotional phrasing. “The booking log recorded 48 unused places” is more auditable than “the analysis uncovered a major attendance crisis.” “The report recommends an eight-week pilot” is clearer than “it is felt that implementation should be considered.” Tentative language is useful when it marks genuine uncertainty, not when it hides who should act.

Revise against the report, not from memory

Write the first version with the final report open. Then run three passes.

Completeness pass

Does the summary state the purpose and the answer?
Are the findings selected for the reader’s decision rather than because they are easy to quote?
Is a required recommendation present?
Is any material limitation omitted?

Accuracy pass

Can every number, group, time period and comparison be found in the report?
Does correlation remain correlation, and does participant feedback remain participant feedback?
Does the conclusion use the same level of confidence as the body?
Does the recommendation stop where the evidence stops?

Brief-and-reader pass

Does the section name and length follow the assignment instructions?
Can the intended reader identify the decision and next action quickly?
Are abbreviations, technical terms and organisational context understandable without reading the full report first?
Does the summary stand alone without adding new citations, tables or evidence that the body never explains?

Cut repeated context before cutting the decisive evidence. A short summary that omits the main result is not concise; it is incomplete.

Use the right kind of revision support

If the summary and report disagree, the task is structural editing: reconcile the claims, evidence and recommendation before polishing sentences. Our guide to academic editing and proofreading explains why that is different from a final language check.

If you want guided feedback on a summary you are developing, academic writing development & revision can help you test the evidence chain and revise your own work while you retain authorship of the assessed submission.