Describing Work Accurately without Inventing Results or Metrics

Accurate descriptions are stronger than impressive claims you cannot explain. Use measured results when you have reliable records; otherwise describe the output, scope, decision or observed change you can support. A percentage is not compulsory. Distinguish what you produced, what someone observed and what you can reasonably attribute to your work.

Build the evidence before polishing the presentation

Create a claim ledger before polishing the resume. For each proposed statement, record the underlying evidence and the strongest wording it supports. “Created a report” is an output. “The team used the report in a planning meeting” is an observed use if you know it happened. “Increased profit” is a causal business claim requiring substantially different evidence.

Check the denominator, time period and comparison behind every figure. If a process took less time once, that does not automatically establish a lasting improvement. If you worked in a team, identify your contribution to the result. Avoid rounding a doubtful estimate into a confident fact merely because the resulting bullet sounds better.

A fictional worked example

Fictional example: Arun reorganised a student club’s registration sheet. He remembers that the next event felt easier to manage, but he did not measure processing time. “Reduced registration time by 40%” would be invented. A defensible statement is, “Reorganised the registration sheet with validation rules and a duplicate-check column for the volunteer team.”

If a volunteer later confirmed that the sheet helped identify duplicate entries, Arun can describe that observed use with appropriate scope. He should not convert it into a claim of eliminated errors or organisation-wide efficiency. The revised description remains concrete and useful because it explains the work and the practical problem it addressed.

Output

Show: A file, feature, analysis or process you actually produced

Check: Does the artefact show the stated contribution?

Observed use

Show: A documented way someone used or reviewed the output

Check: Do you know this happened, or are you assuming it?

Measured impact

Show: A defined change with a valid comparison and attribution

Check: Are the period, denominator and responsibility defensible?

Apply the process to your own material

1. List: Collect the strongest claims you want to make before editing their wording.
2. Trace: Connect each claim to a record, artefact or observation you can explain.
3. Limit: Choose language that matches the evidence and separates individual from team contribution.
4. Review: Remove invented precision and check confidentiality before sharing the final description.

Use an actual piece of your own work for this exercise, or create a clearly labelled practice example. Keep the original and revised versions so you can explain what changed. Ask a reviewer to identify the main contribution without additional explanation, then note where their interpretation differs from what you intended. Revise the evidence or wording that caused the mismatch rather than adding an unsupported claim to make the result sound stronger.

1. List

   Collect the strongest claims you want to make before editing their wording.

2. Trace

   Connect each claim to a record, artefact or observation you can explain.

3. Limit

   Choose language that matches the evidence and separates individual from team contribution.

4. Review

   Remove invented precision and check confidentiality before sharing the final description.

Move forward only when the previous step is clear enough to explain. A finished-looking file still needs a content and usability review.

Review the result before sharing

Look especially closely at verbs such as increased, reduced, transformed, led and owned. Each implies something about causation or responsibility. Replace an unsupported result with a specific action and visible output. Keep a private evidence note for interviews, but do not publish confidential records simply to prove a resume bullet. Accuracy and appropriate disclosure must work together.

A closer check that strengthens the evidence

Check the difference between precision and usefulness. A verified count can still distract if it tells the reader nothing relevant. For example, the number of slides in a presentation may be less useful than the audience and decision the presentation supported. Choose figures that clarify scope or evidence, and leave out counts included only to make the bullet look quantitative. If a result belongs to a wider team or period, explain that boundary. The objective is not to maximise the number of measurable-looking statements; it is to make the contribution easier to understand and harder to misinterpret.

Common questions

What can I say when I have no performance metrics?

Describe the problem, your action, the method and the output. Specific scope can demonstrate skill without a business-impact figure. Explain what was checked or delivered and who it was for. An honest account of useful work is preferable to an unsupported percentage added to imitate a resume example.

Can I estimate a result from memory?

Treat uncertain recollections cautiously and avoid presenting them as exact measurements. If the figure matters, check reliable records or use wording that does not depend on it. Do not invent a precise percentage from a general impression. You should be able to explain how a stated result was established.

Is it acceptable to describe team results?

Yes, if you identify the team context and your contribution. Do not imply sole ownership of a shared achievement. Explain which part you delivered, coordinated or supported. A reviewer can value collaboration while still needing to understand your individual evidence, so make the relationship explicit rather than ambiguous.

What is wrong with saying I improved a process?

The phrase may be accurate, but it needs context. Explain what changed and what evidence supports calling it an improvement. If you only proposed a change, say proposed. If it was implemented but not evaluated, describe the implementation. Avoid collapsing recommendation, delivery and measured benefit into one unsupported claim.

Can an AI tool add stronger metrics to my resume?

It should not invent facts. You can use a tool to improve clarity, but verify every statement against your records. Reject numbers, titles or outcomes that were not supplied and supported. The final application is your account of your work, so responsibility for accuracy remains with you.

Should I remove all numbers to be safe?

No. Verified numbers can clarify scope, such as the number of records checked or sessions coordinated. Use them when they are meaningful and accurately defined. The problem is unsupported precision, not numbers themselves. Avoid adding irrelevant counts that make the bullet longer without helping a reader understand the contribution.

How do I discuss confidential achievements?

Use an approved level of generality and obtain permission where needed. Explain your role and method without exposing sensitive records, client identities or internal figures. If a claim cannot be shared appropriately, choose another example. A portfolio should not create a confidentiality problem in the process of demonstrating competence.

How can I correct an exaggerated draft?

Return to the evidence and rewrite the claim at the strongest defensible level. Replace invented impact with actual work, clarify team ownership and remove unsupported titles. Check related profiles and portfolio pages for the same wording. Consistency matters, so correcting only one document may leave the misleading claim elsewhere.

Sources and further reading

Reviewed 7 November 2026. Examples and exercises are original editorial guidance. University advice reflects its own context; employer instructions and local requirements take priority.

• MIT Career Advising — Resumes
• Harvard Mignone Center — Resume and cover-letter guidance
• Yale Office of Career Strategy — Technical resume sample