How to Write an Analysis (With Examples and Tips)

What is an analysis?

An analysis is a reasoned interpretation supported by evidence. It moves from observable information to an insight, judgment or recommendation. Depending on the purpose, it may compare alternatives, identify causes, assess effectiveness, explain themes or forecast likely outcomes.

A basic analytical unit contains four parts:

1. Claim: The point the writer wants the reader to accept.

2. Evidence: The data, example, quotation or observation supporting the claim.

3. Reasoning: The explanation of how the evidence supports the claim.

4. Significance: Why the conclusion matters for the question or decision.

If a paragraph includes evidence without reasoning, it is closer to a summary or data dump. If it includes a claim without evidence, it is an opinion rather than a complete analysis.

Why analysis is important

Analysis converts information into something a reader can use. A marketing manager may analyze channel performance to allocate a budget. A student may analyze imagery in a novel to explain a theme. An operations team may analyze delays to find the conditions associated with missed deadlines.

The same discipline improves credibility across these contexts: define the question, select relevant evidence, separate observation from inference and acknowledge uncertainty.

Analysis vs. summary

Summary	Analysis
Condenses what happened or what a source says	Explains how, why or with what significance
Emphasizes main facts	Evaluates patterns, causes, implications or relationships
Can be neutral and descriptive	Makes an interpretation or judgment supported by evidence
Answers “What?”	Often answers “How?”, “Why?” or “So what?”

A report can contain both. A short summary may orient the reader before the analysis interprets the evidence.

Common types of analysis

Business analysis

Business analysis examines performance, processes, customers or alternatives to support a decision. It may combine quantitative metrics with interviews, operational context and financial constraints.

Data analysis

Data analysis cleans, organizes and interprets data to identify patterns or test a question. The written result should describe the dataset, method, findings, limitations and implications.

Financial analysis

Financial analysis evaluates revenue, expenses, margins, cash flow, investments or risk. It should state the period, accounting definitions and assumptions used.

Process analysis

Process analysis studies how work moves from one step to another. It can identify delays, rework, capacity limits and control failures before recommending changes.

Comparative analysis

Comparative analysis assesses two or more options using consistent criteria. A fair comparison defines those criteria before declaring one option preferable.

Literary or rhetorical analysis

Literary analysis interprets elements such as language, structure, imagery and character. Rhetorical analysis considers how a creator uses evidence, style and appeals to influence an audience.

Root-cause analysis

Root-cause analysis investigates why an outcome occurred. It distinguishes symptoms from contributing conditions and avoids claiming causation without adequate evidence.

How to write an analysis in 10 steps

1. Clarify the purpose and audience

Determine what decision, question or assignment the analysis must address. Identify who will read it, what they already know and what action they may take. An executive brief and an academic essay can analyze the same topic but require different depth, terminology and evidence.

Rewrite a broad topic as a focused question. “Customer retention” is a subject; “Which factors best explain the decline in first-year customer retention during the second quarter?” is an analytical question.

2. Define the scope

Set boundaries for time period, geography, population, sources and variables. Scope prevents the analysis from making conclusions beyond the evidence.

For example: “This analysis covers U.S. paid-search campaigns from January through June and excludes branded search.” That sentence helps readers interpret the result correctly.

3. Gather credible evidence

Collect the information needed to answer the question. Sources might include internal records, interviews, surveys, experiments, peer-reviewed research, official statistics or primary documents.

Record where each item came from, when it was collected and any known limitations. Do not select only evidence that supports your preferred conclusion.

4. Evaluate evidence quality

Consider relevance, accuracy, sample size, definitions, possible bias and whether the source directly supports the claim. A large number of social-media comments may reveal themes but does not automatically represent the entire customer population.

Distinguish correlation from causation. If two metrics change together, additional evidence may be required to conclude that one caused the other.

5. Organize the information

Group evidence by theme, time, cause, comparison criterion or stage of a process. Tables, charts and coding categories can reveal patterns that are hard to see in raw notes.

Create an outline that matches the reader's decision. A useful workplace structure is: question, evidence, findings, interpretation, alternatives, recommendation and limitations.

6. Develop a central conclusion or thesis

State the main interpretation in one or two sentences. It should answer the analytical question and be specific enough to test against the evidence.

Weak: “The campaign had mixed results.”
Stronger: “The campaign increased low-cost traffic but reduced lead quality because growth came primarily from broad informational queries.”

The stronger version identifies both the observed outcome and the explanation that the analysis must support.

7. Write the introduction

Orient the reader without delaying the point. A concise introduction normally includes:

• The question or problem

• The scope and relevant context

• The central conclusion

• A brief preview of the structure

For a short business analysis, the conclusion may appear in the first paragraph. Academic assignments may require a thesis at the end of the introduction.

8. Build analytical body paragraphs

Give each paragraph one main purpose. Start with a claim, present evidence, explain the connection and state the significance.

Example structure:

Claim: Mobile checkout friction appears to be the largest controllable source of abandonment.
Evidence: Mobile users represented 71% of checkout sessions but 84% of payment-page exits. Session reviews also showed repeated address-field errors.
Reasoning: Because exits are disproportionately concentrated on mobile and coincide with a specific form problem, the evidence points to usability rather than product demand alone.
Significance: Fixing the form should be tested before increasing acquisition spend.

9. Address alternatives and limitations

Explain other plausible interpretations and why the evidence supports or does not support them. Note missing data, measurement changes, small samples and assumptions.

Limitations do not necessarily weaken good analysis. They show the reader where the conclusion is strong, where it is tentative and what evidence would improve confidence.

10. Conclude with implications or action

Restate the conclusion in light of the evidence rather than repeating the introduction word for word. Explain what the reader should understand, decide or investigate next.

Recommendations should follow directly from the findings. Assign an owner, timing or measurement plan when the analysis is intended to drive action.

Analysis structure template

Use this outline for a report, memo or essay:

Title: [Specific subject and analytical focus]

Question or purpose:
[What the analysis evaluates and why]

Scope and method:
[Time period, population, sources, criteria and method]

Central conclusion:
[The main answer in one or two sentences]

Finding 1:
[Claim, evidence, reasoning and significance]

Finding 2:
[Claim, evidence, reasoning and significance]

Alternative explanations:
[Other interpretations considered]

Limitations:
[Data or method constraints]

Recommendation or implication:
[Action supported by the findings]

Business analysis example

Topic: Why customer-support resolution time increased

Purpose and scope: This analysis examines support tickets closed from April through June to identify the main factor associated with an increase in median resolution time from 18 to 27 hours.

Analysis: Ticket volume rose 9%, but the increase alone does not explain the 50% change in median resolution time. Product-integration tickets represented 22% of volume yet accounted for 48% of tickets open longer than 48 hours. Interviews with support staff showed that these cases required manual escalation to engineering and lacked a defined ownership path. By comparison, billing and account-access tickets maintained response times close to the previous quarter despite similar volume growth.

Conclusion: The evidence indicates that unresolved integration escalations, rather than general ticket volume, were the primary source of delay. The support and engineering teams should test a shared escalation queue with severity definitions and measure median time to owner, time to first engineering response and total resolution time for four weeks.

Limitations: The dataset covers one quarter and does not measure the technical complexity of every ticket. The pilot should therefore confirm that faster routing improves resolution without increasing reopens.

Literary analysis example

Topic: Repeated doorway imagery in a short story

Thesis: The story's repeated doorway imagery represents the protagonist's hesitation between preserving a familiar identity and accepting an uncertain future.

Analysis paragraph: Doorways appear at each major decision, but the protagonist initially observes them without crossing. This repeated physical pause turns an ordinary setting into a visual pattern of indecision. In the final scene, the character leaves the door open after stepping outside, suggesting that change does not require rejecting the past completely. The shift from watching to crossing supports the interpretation that the conflict is about willingness to act rather than the absence of fear.

The paragraph does not merely mention the images. It connects repetition, action and change to the thesis.

Tips for writing a strong analysis

Use precise language

Replace vague claims such as “a lot,” “better” or “successful” with defined measures. State the period, population and comparison behind each figure.

Separate facts from interpretation

Signal the difference between what the evidence directly shows and what you infer. Phrases such as “the data shows,” “this pattern suggests” and “one possible explanation is” communicate different levels of certainty.

Explain every important piece of evidence

Do not assume a quotation or statistic proves the point by itself. Add the reasoning that connects it to the conclusion.

Use consistent comparison criteria

When evaluating options, assess each one against the same dimensions. Changing criteria halfway through creates a biased comparison.

Prefer relevant evidence over volume

Three strong sources that directly address the question are usually more useful than many loosely connected facts. Remove evidence that does not advance the analysis.

Edit for decision value

Ask what the reader should know or do after each section. Move supporting detail to an appendix if it interrupts the main argument.

Verify citations and calculations

Check source attribution, units, percentages, denominators and chart labels. Recalculate important results and confirm that quoted material remains in context.

Common analysis mistakes

• Repeating source material without interpreting it

• Starting with a conclusion and ignoring contradictory evidence

• Confusing correlation with causation

• Comparing inconsistent time periods or populations

• Using undefined metrics

• Making a recommendation unrelated to the findings

• Hiding important limitations

• Introducing new evidence only in the conclusion

• Using complex language to disguise weak reasoning

Dokie can help turn a completed analysis into a clear presentation for decision-makers. Users can organize the question, evidence, findings, limitations and recommendations into an editable slide narrative, making it easier to show how each conclusion follows from the underlying material.

Dokie also supports PowerPoint export and slide-level revision for business reviews, research summaries and academic presentations. Verify every figure, quotation, chart and citation against the original sources, and remove confidential or restricted data before uploading any material.

Frequently asked questions

What is the first step in writing an analysis?

Define a focused question and the audience's decision. A clear question determines what evidence is relevant and prevents the analysis from becoming a general description.

How long should an analysis be?

It should be long enough to support the conclusion and address material limitations. A management memo may be one page, while an academic or technical analysis may require several sections and appendices.

Can an analysis use first person?

It depends on the context and style guide. First person may be appropriate when describing your method or judgment, but the conclusion still needs evidence.

What is the difference between evidence and analysis?

Evidence is the information observed or collected. Analysis is the reasoning that explains what the evidence means, how pieces connect and why the result matters.

Does an analysis need a thesis?

Most analyses benefit from a central conclusion or thesis. In exploratory work, the conclusion may emerge after reviewing evidence rather than being fixed in advance.

How do you make an analysis objective?

Define criteria before comparing options, use credible sources, consider contrary evidence, separate observation from inference and state limitations clearly. Complete neutrality may be impossible, but transparent methods improve fairness.