All guides

    For founders & builders

    How to Use GitHub in Your Performance Review Without Counting Commits

    Learn how to turn pull requests, reviews, and closed issues into clear performance-review examples without counting commits or sharing source code.

    Bloomly EditorialAugust 19, 2026 9 min read

    When review season starts, GitHub can help you remember what you worked on. But a list of pull requests is not a self-review. Your manager still needs to know why the work mattered, what part you owned, and what changed after it shipped. This guide shows you how to turn everyday GitHub activity into clear career stories. You will also see how Bloomly can help without reading your source code or diff patches. You do not need to become a data analyst to do it.

    What this guide covers
    1. 01Your commit count does not tell the story
    2. 02Use four questions for every important piece of work
    3. 03Pull requests can show more than coding
    4. 04Reviews and comments can show leadership
    5. 05Closed issues can bring forgotten work back
    6. 06You do not need to share your code
    7. 07A 10-minute Friday habit is enough
    8. 08Turn the note into a self-review sentence
    9. 09Give the team credit without hiding your part
    10. 10Common mistakes

    Your commit count does not tell the story

    Imagine two engineers. One makes 80 commits during a large cleanup. The other makes a small fix that stops customers from being charged twice. The first person did more visible activity. The second may have created more value. That is why commit totals and lines of code make poor review material.

    A strong review is about change. Did the product become faster? Did the team avoid an outage? Did a launch move forward because you solved a hard problem? Did your feedback help another engineer ship safer work? GitHub can point you to the moment. You add the meaning.

    Use four questions for every important piece of work

    You do not need a complex system. Open a pull request, review, or issue that seems important. Then answer four questions in plain language.

    • What was the problem? Explain what was broken, slow, risky, or confusing.
    • What did you do? Name the choice you made or the part you owned.
    • Who did it help? Think about customers, coworkers, or another team.
    • What changed? Add a number if you have one. If not, describe a clear result.

    Here is a simple example. The GitHub record says, “Merged PR #4821: retry failed payments.” Your note could say, “I found that temporary network errors were causing some customers to submit payment twice. I added a safe retry flow, worked with Support to test the common cases, and helped cut duplicate-payment tickets from eight a week to one.” The second version tells a manager what the work meant.

    If you do not have the final result yet, do not guess. Write what you expect to improve and add a reminder to check later. For example: “Shipped the new retry flow on June 14. I will check payment errors after two full weeks.” Honest follow-up is stronger than a made-up number.

    Pull requests can show more than coding

    A pull request can remind you about the code, but it can also remind you about judgment. Maybe you cut the scope so the team could ship before a deadline. Maybe you added a rollout flag because the change was risky. Maybe you brought Design and Support into the decision before launch.

    Those choices often matter more than the size of the change. When you look back at a pull request, scan the description and conversation. Ask what made the work hard. Then write one or two sentences about the tradeoff you handled.

    • Instead of “Built a new settings page,” write “Cut the first version to three key controls so the team could launch before the billing change.”
    • Instead of “Refactored the upload service,” write “Moved uploads behind a safe rollout flag and avoided downtime for existing customers.”
    • Instead of “Fixed onboarding,” write “Removed the step causing most drop-offs and raised completion from 54% to 68%.”
    • Instead of “Worked on migration,” write “Planned the move in two stages, gave every team a rollback path, and finished without an outage.”

    These are sample sentences, not claims to copy word for word. Replace the project, numbers, and choices with your own. The best review language sounds like your real work because it is based on your real work.

    Reviews and comments can show leadership

    Engineers often forget the value of code review because they did not write the final code. But a thoughtful review can prevent a bug, improve a design, or help a teammate learn a new pattern. That is real work.

    Do not report that you left 120 comments. Pick the comments that changed the outcome. Maybe you spotted a privacy issue before launch. Maybe you suggested a smaller design that saved a week. Maybe you helped a new teammate finish a difficult change.

    • “Caught a missing permission check during review and helped the team fix it before release.”
    • “Suggested reusing the existing service instead of building a second one, which kept the project on schedule.”
    • “Worked through the first payment change with Maya. She used the same pattern on two later projects without help.”
    • “Asked the team to add a rollback plan before launch. We used it when the first rollout showed an error spike.”

    The useful part is not that you commented. It is that your comment improved the work or helped another person. Name that result.

    Closed issues can bring forgotten work back

    Issues often hold work that never became a large pull request. You may have researched a problem, made a call, closed an old process, or helped the team stop doing work that was no longer useful. These decisions belong in a review when they saved time or reduced risk.

    For example, an issue might say, “Close old export job.” Your review note could say, “I found that the old export job was still waking the on-call engineer twice a month. I removed it, updated the runbook, and we had no related pages during the next eight weeks.” That is easy to understand, even for a manager who never opened the issue.

    Canceled work can count too. Stopping a weak idea before the team spends a month on it shows good judgment. Explain what you learned, why you stopped, and what the team did instead.

    You do not need to share your code

    It is reasonable to be careful with work code. A career journal should not copy private source code, customer data, or secret plans. Most useful career details do not require any of that.

    A safe note can include the project name, your role, the decision, an approved number, and a private link back to the original work. If a number is sensitive, use a percentage, a range, or a simpler result such as “cut weekly support tickets by more than half.” Follow your employer's rules before saving or sharing anything.

    Bloomly's GitHub connection is read-only. For repositories you choose, it can use your pull requests, reviews, and closed issues to make private work records. It may use text you wrote in a pull request, a limited set of commit titles tied to that pull request, and total change counts. It does not read source code or diff patches. It also does not pull standalone commits or releases.

    A 10-minute Friday habit is enough

    At the end of the week, open the work you merged, reviewed, or closed. Skip routine updates. Choose one to three items that changed a result, solved a hard problem, or helped someone else move faster.

    • Save the link and date.
    • Write the problem in one sentence.
    • Write what you owned in one sentence.
    • Add the result now, or set a date to check it later.
    • Note any teammate who helped so the story gives fair credit.

    This small habit matters because details fade quickly. Months later, you may remember that a launch was hard but forget the exact problem you solved. A short note made on Friday keeps the useful part.

    Turn the note into a self-review sentence

    When it is time to write your review, lead with the result. Then explain your action and scope. Add the GitHub link for your own records or wherever your company allows it.

    • Result: “Reduced failed checkout attempts by 31%.”
    • Action: “I traced the issue to a retry bug and built a safer payment flow.”
    • Scope: “I worked with Payments, Support, and SRE to test and roll it out.”
    • Growth: “I later helped two other teams use the same pattern.”

    Put together, that becomes: “I reduced failed checkout attempts by 31% by finding a retry bug and building a safer payment flow. I worked with Payments, Support, and SRE on the rollout, then helped two other teams use the same pattern.” Clear, specific, and easy to believe.

    Give the team credit without hiding your part

    Most useful software work is shared. A designer may shape the flow. A product manager may set the goal. SRE may make the rollout safer. Saying “I did everything” will not sound true. Saying only “we did it” can make your own work disappear.

    Use both. Start with the team result, then name the part you owned. For example: “Our checkout team cut failed payments by 31%. I found the retry problem, built the safer flow, and led testing with Support and SRE.” This shows teamwork and responsibility at the same time.

    If you are unsure how much credit to take, ask a teammate to read the sentence. They can tell you whether it feels fair and whether you missed an important part of your work. A clear, honest sentence is more convincing than a dramatic one.

    Common mistakes

    • Listing totals. A count of commits or reviews does not explain why they mattered.
    • Using vague words. Replace “helped with” with the choice you made and the result.
    • Taking all the credit. Name partners and make your own part clear.
    • Adding numbers you cannot support. A smaller true claim is better than a big guess.
    • Saving code or private customer details. Keep the career story, not the sensitive material.
    • Waiting until review week. Ten minutes now can save hours of searching later.

    GitHub is a useful memory aid, not a scorecard. Start with the pull request, review, or closed issue. Then add the part GitHub cannot know: the problem, your choice, the people you helped, and the result. A few clear stories will do more for your review than a giant list of commits. Save those stories while the details are fresh, and your next self-review will feel much easier.

    Your next ten minutes

    Put this into practice

    1. 1

      Choose one moment

      Start with a project, decision, or result that still feels important.

    2. 2

      Add what changed

      Write down your part, why it mattered, and any proof you can return to.

    3. 3

      Save it while it is fresh

      A useful note today is better than a perfect memory six months from now.

    Quick answers

    Frequently asked questions

    Should I include my commit count in a performance review?

    Usually, no. Commit totals show activity, not value. Pick a few important pull requests or issues and explain the problem, your role, and the result instead.

    What if my GitHub work does not have a clear number?

    Use a clear result you can support. You may have avoided a risk, helped another team launch, made on-call easier, or shortened a process. Add a number later if one becomes available.

    Do code reviews belong in a self-review?

    Yes, when your feedback changed the work. Describe the bug you caught, the decision you improved, or the teammate you helped. Do not rely on the number of comments alone.

    Can Bloomly help without reading source code?

    Yes. Bloomly can use read-only information about your pull requests, reviews, and closed issues from repositories you choose. It does not read source code or diff patches.

    How often should I save GitHub wins?

    A quick weekly check works well. Choose one to three meaningful items, save the link, and write why each one mattered while you still remember the details.

    Sources

    Claims in this article are backed by the following published sources.

    1. Ebbinghaus, H. (1885). Memory: A Contribution to Experimental Psychology. Read

      Original research on the forgetting curve — the basis for claims that most work memory degrades within weeks, motivating the case for a contemporaneous career log.

    2. Cappelli, P. & Tavis, A. (2016). The Performance Management Revolution. Harvard Business Review. Read

      Industry-defining piece on the shift from annual reviews to continuous feedback at Adobe, Deloitte, Microsoft, GE, and others.