# OKR key results that measure real outcomes: how to fix vanity metrics on Mac

> Vanity metrics hide in OKRs when key results count activity. Learn how to write outcome key results, score them honestly and fix them on your Mac.

URL: https://www.lazur.app/for/product-managers/problems/okr-key-results-real-outcomes
Updated: 2026-11-08

OKR key results that measure real outcomes describe a change in customers, revenue, or behavior. They do not describe the work your team does. A key result that counts tasks closed or launches shipped is a vanity metric, and it can hide a quarter where nothing important moved. This guide shows how to spot those lines, rewrite them, and score them without fooling yourself.

## Quick answer

- A vanity metric counts output or activity. It can look strong while the customer or the business stays flat.
- Write each key result as a change, with a baseline, a target you can defend and a date.
- A high score on output lines proves the work got done, not that the goal moved.
- When a number becomes a target, people optimize it. That is Goodhart's law in action.
- Lazur explains what is on screen, points at the field you name and drafts wording. It does not verify your numbers.

## What vanity metrics in okrs costs you

- **Quarters spent on work that moves nothing.** The first cost is time. Asana says on its own marketing page that 60% of a person's time at work goes to work about work, not skilled work. The page cites its Anatomy of Work Index and more than 10,000 knowledge workers, and it gives no survey year. Treat it as a vendor figure, not a neutral study. A team that tracks output can feel busy while the outcome stays flat.
- **Roadmaps filled with the wrong work.** A vanity metric tells the team what to fund. If a key result counts launches, the roadmap fills with launches. Features that would move a customer measure get pushed out of the quarter. Nobody decides this on purpose. The OKR made the decision for them.
- **Scores that mislead leadership.** Leaders read a high score as success. When a team scores well on output, the review ends early. The problem shows up later, in a churn chart or a support queue, long after the quarter closed. By then the team has a clean record and a weak result.
- **Trust and morale.** People notice when a metric is cosmetic. The team learns that the number matters more than the customer, and some people stop reporting bad news early. The WHO describes burn-out as an occupational phenomenon from chronic workplace stress that has not been managed. No source here links vanity metrics to burn-out directly, so treat any link as a hypothesis.
- **Your own record at review time.** A product manager who reports output for several quarters has a thin story at review time. Outcomes are the proof of the work. Without them, the record shows activity, which is hard to defend when priorities change.

## Why it happens

1. **Activity is easy to count.** Tasks, tickets, and launches already sit in your tracker. Outcomes need a second source, such as analytics, a survey, or support data. The easy number wins the first draft, and nobody goes back to change it.
2. **Key results drift toward to-do lists.** Teams often write the work into the key result because the work is what they own. Google's guide is clear that key results should describe outcomes, not activities. Even so, a first draft often ends up as a to-do list with a score attached.
3. **Goodhart's law: a target changes behavior.** When a number becomes a target, people optimize it. Manheim and Garrabrant's paper says optimizing a system against a metric can cause failures. Fire and Guestrin found that publication counts rose while the papers changed shape. OKRs are a target system, so the same pressure applies.
4. **OKR scores used for pay or ratings.** When OKR scores feed bonuses or performance ratings, people protect the score. Google says OKRs are not synonymous with performance evaluation. A team that believes the scores decide pay will choose safe, easy key results.
5. **Goals set too low to miss.** Google's guide warns that scoring higher may mean the aspirational goals are not being set high enough. A safe target gives a high score and little learning. Teams then pick activity they can finish, because that is the thing most likely to score well.
6. **No baseline, so no proof.** If nobody measured the starting value, any outcome looks like progress. The team picks a number it can reach and calls the result a win. Without a baseline, an outcome key result turns back into output with extra steps.

## How to fix it

- **Rewrite each key result as a change.** Start with the sentence the customer or the business would care about. Write the change first, then the measure. A clear format is the thing that changes, the measure, the baseline, the target, and the date. A plain doc or sheet is enough, and nothing needs to be bought.
- **Pair every output with an outcome.** Keep the launch or build line if it is a real dependency. Add an outcome line that proves the work helped. If the outcome misses, the output still happened, and you learn that the output was not enough on its own.
- **Run the what-changed test in review.** In the review, read each key result aloud and ask what changed for the customer. If nobody can answer, mark the line as output. Do not change the score yet. Rewrite the line first, then score it.
- **Set a baseline before a target.** Pull the starting value from data you already have. Write the date you measured it next to the number. Set the target after the baseline exists, and explain it in one sentence that a new hire would understand.
- **Report outcomes before scores.** Show the outcome line before the score. Ask whether the outcome moved, and by how much. Then read the score as a planning record. Keep OKRs out of individual performance ratings, as Google's guide advises.
- **Add a counter-metric to every outcome.** For each outcome, name a measure that would show damage if the team pushed the first number too hard. If the outcome is conversion, watch refunds or support contacts. The counter-metric is the early warning that a number is being gamed.
- **Review mid-quarter and cut what does not move.** Check each outcome every few weeks. If an outcome has not moved after honest work, change the plan or retire the line. Write the reason in the doc, so the next quarter starts from evidence and not from habit.
- **Test the wording with someone outside the team.** Show the rewritten key result to a designer, an engineer, or a support lead. Ask what they think it measures. If their answer differs from yours, keep rewriting until the answers match.

![Product manager reviewing OKR key results on a laptop at a desk](https://images.lazur.app/content-images/problem/okr-key-results-real-outcomes/okr-key-results-that-measure-real-outcomes-1-37a80a.jpg)

## What counts as a vanity metric in an OKR

A vanity metric is a number that looks good on a slide and tells you little about the product or the business. In an OKR, it usually appears as a key result that counts output. Output is work you finished, such as tickets closed, screens designed, experiments launched, or meetings held. Output is not bad. The trouble starts when output takes the place of the outcome, so the team reports progress that nobody outside the team can feel.

| Output key result (weak) | Outcome key result | What would change |
| --- | --- | --- |
| Launch the new onboarding flow | New users finish setup in their first session, at the target the team agreed on | Users complete the job the flow was built for |
| Publish the planned help articles | Setup tickets fall to the agreed level | Fewer people get stuck, which is the reason the articles exist |
| Run the scheduled user interviews | The team can name the top reason trial users leave, backed by evidence | A product decision changes because of what users said |
| Ship the dashboard | Weekly planning starts from the dashboard instead of a spreadsheet export | A recurring decision uses better data |

_The same work, written as output and as outcome. Set the targets yourself._

Read the first column aloud, then ask who would notice if that work stopped. Read the second column and ask what a customer or the business would feel. If the answer to the first question is nobody outside the team, the line is probably output. Google's re:Work guide puts it plainly: key results should describe outcomes, not activities.

## A quick test for any key result

1. Ask what changes for a customer or the business if this number hits its target.
2. Ask whether the number could rise while the business gets worse. If it could, the number is a weak signal.
3. Ask whether you know the starting value. No baseline means no proof of change.
4. Ask who outside the team would notice if the work stopped.
5. Ask what you would do differently if the number missed. If the answer is nothing, drop the line.

The second question catches most vanity metrics. A count of published help articles can rise while setup tickets rise too. The article count looks healthy, but the customer still gets stuck. Pair every output line with an outcome that would show whether the output helped.

## When an output key result is fine

Some output lines are honest. A launch tied to a contract, a platform cutover, or a legal deadline is a fact the business must meet. Keep that launch as a dependency, not as the only result of the quarter. Write the launch line, then add the outcome line that explains why the launch mattered. Clear labels stop anyone from mistaking a milestone for the whole goal.

## How to write a baseline and a target you can defend

A baseline is the value of the outcome before the quarter starts. Pull it from a tool you already use, such as an analytics dashboard or an export into Google Sheets. Write the date you measured it next to the number. A target should be a number you can explain in one sentence. Round numbers are tempting. A target you cannot justify falls apart at the first review.

![Product team in a planning meeting checking goals on a screen](https://images.lazur.app/content-images/problem/okr-key-results-real-outcomes/okr-key-results-that-measure-real-outcomes-2-23f826.jpg)

New features often have no baseline. In that case the first key result can be an output that creates one, such as adding the tracking event or running a short measurement window. Label it honestly as output. Set the outcome target after the baseline exists, not before.

> **Find the baseline first:** Before you rewrite a key result, find its current value and the date you measured it. If you cannot find either in an afternoon, that gap is a finding. Write Not measured in the line and make the baseline the first step.

## How scoring can hide a vanity metric

Google's guide grades OKRs on a 0.0 to 1.0 scale. It says the sweet spot is somewhere in the 60 to 70% range. It also warns that scoring higher may mean the aspirational goals are not set high enough. Vanity metrics make this worse. A team that scores 1.0 on output lines has finished its list. That score says nothing about whether a customer got anything.

- **60%** Share of work time Asana says goes to work about work, not skilled work, from its own marketing page ([source](https://asana.com/resources/why-work-about-work-is-bad))
- **60 to 70%** Google re:Work's suggested sweet spot for OKR scores graded on a 0.0 to 1.0 scale ([source](https://rework.withgoogle.com/intl/en/guides/set-goals-with-okrs))
- **120.7 million** Papers analyzed in a GigaScience study of metric distortion in academic publishing ([source](https://pmc.ncbi.nlm.nih.gov/articles/PMC6541803/))
- **31 Dec 2025** Date Microsoft retired Viva Goals, its OKR tool ([source](https://learn.microsoft.com/en-us/viva/goals/goals-retirement))

Read the score and the outcome as two separate facts. Ask whether the outcome moved, and how far, before you celebrate a high score. In a review, read the outcome lines first. If every line is output, the score is a planning record, not a result.

## Goodhart's law and why good numbers go bad

Goodhart's law, in plain words, says a measure stops being a good measure once people are rewarded for hitting it. Manheim and Garrabrant's 2018 paper on variants of the law says that optimizing a system against a metric can cause failures. It refers to at least four mechanisms behind the problem. You do not need the full taxonomy to spot the risk. You need to ask what the number makes people do.

Fire and Guestrin's 2019 study in GigaScience analyzed 120.7 million papers. Its conclusion was that the number of publications has ceased to be a good metric. Over the same period, mean authors per paper rose from 1.41 in 1900 to 4.51 in 2014, for papers with at least five references. Mean paper length fell from 14.4 pages in 1950 to 8.4 pages in 2014. The study is observational and covers academic publishing, not OKRs. It still shows how a count can bend behavior.

## What to do if your OKR tool has been retired

Microsoft retired Viva Goals on 31 December 2025. It had stopped adding new features as of 5 December 2024. The retirement page listed export options through an API, Excel, and PowerPoint, and it recommended exporting before the retirement date. If your team never exported, ask your admin whether a copy exists. The Microsoft pages do not describe a recovery path after retirement, so do not count on one.

A dedicated app is not required. A shared doc, a sheet, or a Notion database can hold OKRs. Use one column for the outcome, one for the baseline, one for the target, and one for the current value with its date. A reader can then see which lines are outcomes at a glance. For shortcuts and workflow tips, see [Notion for product managers on Mac](https://www.lazur.app/for/product-managers/notion).

## Mac apps for OKR tracking and what they cost

| App | Mac requirement | Price as listed | Version and ratings |
| --- | --- | --- | --- |
| Profit OKR Software | macOS 10.15 or later | Free | Version 3.2.12, no ratings yet |
| OrderOkr, OKR Goals Management | macOS 15 or later | Free with in-app purchases, from $2.99/month or $24.99 lifetime | Version 3.5.2, 3.4 of 5 from 5 ratings |

_Two Mac App Store listings. Prices change, so confirm on each listing._

Neither listing has enough ratings to judge quality. Read the feature list, then test one real quarter before you pay. Several other OKR apps in search results looked iPad first, so check the Mac requirement before you install. A plain spreadsheet or document does the same job for a small team.

## Where Lazur fits, and where it stops

Lazur is a voice assistant for the Mac. Hold ⌥⌘ to ask about anything on screen, and Lazur answers out loud. Ask where a column is and Lazur draws a circle and a label around it. Ask how to do something and Lazur guides you one click at a time, waiting for each click. Hold fn to dictate clean text at the cursor in any app, including a [Linear](https://www.lazur.app/for/product-managers/linear) issue.

Lazur does not click buttons, edit your OKR tool, or change your data. It cannot confirm that a number is right. Open the source, read the query, and decide whether the target is fair. Use Lazur to explain a metric in plain words, point at the field that drives it, or draft a clearer outcome line that you then edit. There is a free trial, and the details are on the [pricing page](https://www.lazur.app/pricing).

## Data rules before you paste OKR numbers into an AI tool

> **Follow your employer's data rules:** OKRs can hold customer names, health details, payment data, or internal financials. Check your employer's rules before you share a screenshot or text with any AI tool, Lazur included. If your job covers client or patient data, such as law, nursing, or accounting, follow those rules first. Lazur looks at the screen only while you hold the shortcut. Audio and the screenshot go to AI providers to answer, and they are not stored. Content is never used to train AI models.

Redact before you ask. Replace names with roles, and blur account numbers in your own notes. Keep the real figures in the source system, where your employer's access rules apply.

## A worked example from trial to paid

Take a team that wants more trial users to start a paid plan. An output key result says: ship the paywall redesign. An outcome version says: trial users who reach the paywall start a paid plan at the agreed rate, measured against last quarter's baseline. The first line is done when the code merges. The second line tells you whether the redesign worked, and it may still miss.

## Mistakes people make when they fix vanity metrics

- Swapping words but keeping the same measure. Output with new wording is still output.
- Adding outcome lines and keeping every activity line. The team tracks more and learns nothing new.
- Choosing an outcome the team cannot move at all. Pair it with a measure the team can influence.
- Checking only at the end of the quarter. Look at the outcome every few weeks so a miss can change the plan.
- Using OKR misses as a performance review. Google says OKRs are not synonymous with performance evaluation.
- Setting targets to protect the score. A safe target gives a high grade and very little information.

## What to do after you rewrite the key results

![Person typing on a Mac keyboard while writing key results](https://images.lazur.app/content-images/problem/okr-key-results-real-outcomes/okr-key-results-that-measure-real-outcomes-3-d8cba0.jpg)

Share the rewritten OKRs with your manager and with one person who does the work every day. Ask them what the outcome line means in their own words. If their version differs from yours, the sentence is not clear yet. Set a review date inside the quarter. At that review, decide whether to keep, change, or drop each outcome, and write down why.

## With Lazur

Lazur is a voice assistant for the Mac that lives in the notch. Hold ⌥ ⌘ and ask; it sees your screen, answers out loud, points at buttons and walks you through apps. Hold fn to dictate. Free for 3 days, then $18/month: https://www.lazur.app/download

- “Where is the baseline column in this OKR sheet?”: You find the column in a second, without scanning every tab. (Points at it)
- “What does this column measure, and what does it miss?”: Lazur answers out loud about what is on screen, so you hear the metric explained while you keep working.
- “Where is the baseline date in this sheet?”: Lazur draws a circle and a label around the cell or column you asked about.
- “Walk me through adding a column for the baseline date.”: Lazur guides you one click at a time and waits for each click, so you stay in control of the file.
- “Draft an outcome version of this key result.”: Lazur drafts the wording as text. You check the target and the source before you use it.
- “Dictate the new key result into this Linear issue.”: Dictation lands at the cursor in any app, so you skip retyping the sentence.
- “Find public examples of outcome key results from product teams.”: Lazur runs the search in its own Chrome tabs and returns a summary and a report.

## A plan you can start this week

- **Day 1: Pull the current OKRs into one list.** Copy every key result into one sheet or doc. Add a column that tags each line as output or outcome. Be honest with the tags. Do not fix anything yet, because a full list makes the pattern visible.
- **Day 2: Find a baseline for each outcome.** For each outcome line, find the current value and the date it was measured. Write Not measured if you cannot find one. That gap is a finding, and you should say so in the next review.
- **Day 3: Run the what-changed test.** Read each line aloud and ask what changed for a customer or the business. Mark the lines that fail as output. Keep the output lines that protect a real dependency, and label them as dependencies.
- **Day 4: Rewrite the weakest lines.** Pick the weakest few lines and rewrite each one as an outcome with a baseline, a target, and a date. Test the wording with one person outside the team. Use Lazur to explain or draft, then check the numbers yourself.
- **Day 5: Add counter-metrics and a review date.** For each outcome, name one counter-metric that would show damage. Put a review date on the calendar inside the quarter. Share the list with your manager before the review, so the first conversation is about evidence.

## FAQ

### What is a vanity metric in an OKR?

A vanity metric in an OKR is a key result that looks like progress but does not show a change in a customer or business state. Most vanity metrics count output, such as launches, tasks, or documents. They are not always wrong. They become a problem when they replace the outcome.

### Are activity key results always wrong?

No. A launch tied to a contract or a regulatory date can be a fair key result. Keep it as a dependency, and pair it with an outcome line that says why the launch matters. Do not let the activity line stand alone for the whole quarter.

### What is a good score for an OKR?

Google's re:Work guide grades OKRs on a 0.0 to 1.0 scale and says the sweet spot is somewhere in the 60 to 70% range. It also warns that a higher score may mean the goals were not ambitious enough. A good score does not prove a good outcome, so read the outcome line too.

### Does Google say OKRs should feed performance reviews?

No. Google's re:Work guide says OKRs are not synonymous with performance evaluation. Many teams still link them anyway. Treat that link as a risk to manage, not as a default setting.

### What is Goodhart's law in simple terms?

When a measure becomes a target, people change what they do to hit it, and the measure stops meaning what it used to. Manheim and Garrabrant's 2018 paper on variants of the law describes at least four mechanisms. Fire and Guestrin's 2019 study shows a version of the problem in academic publishing.

### Can Lazur check whether my OKR numbers are correct?

No. Lazur can explain what is on screen, point at the field you name, and draft wording. It does not verify your data. Check the number at the source, then decide whether the target is fair.

### Which OKR apps run on a Mac?

Profit OKR Software and OrderOkr both list macOS support on the Mac App Store. Profit OKR needs macOS 10.15 or later and is free. OrderOkr needs macOS 15 or later and offers in-app purchases. Check each listing for current prices, and test with one real quarter before you commit.

### Is it safe to paste OKRs with customer data into an AI tool?

Check your employer's data rules first. If your work covers client or patient data, such as in law, nursing, or accounting, follow those rules strictly. Lazur sends audio and screenshots to AI providers only to answer, and they are not stored, but your employer's policy still decides what you can share.

## Sources

1. [Google re:Work: Set goals with OKRs](https://rework.withgoogle.com/intl/en/guides/set-goals-with-okrs), Google re:Work, checked 2026-11-08
2. [Viva Goals retirement | Microsoft Learn](https://learn.microsoft.com/en-us/viva/goals/goals-retirement), Microsoft, checked 2026-11-08
3. [Over-optimization of academic publishing metrics: observing Goodhart's Law in action](https://pmc.ncbi.nlm.nih.gov/articles/PMC6541803/), GigaScience / PubMed Central, checked 2026-11-08
4. [How Work About Work Gets in the Way of Real Work](https://asana.com/resources/why-work-about-work-is-bad), Asana, checked 2026-11-08
5. [Burn-out an 'occupational phenomenon': International Classification of Diseases](https://www.who.int/news/item/28-05-2019-burn-out-an-occupational-phenomenon-international-classification-of-diseases), World Health Organization, checked 2026-11-08
6. [Categorizing Variants of Goodhart's Law](https://arxiv.org/abs/1803.04585), arXiv, checked 2026-11-08
7. [Profit OKR Software: Mac App Store](https://apps.apple.com/us/app/profit-okr-software/id1489934575?mt=12), Apple App Store, checked 2026-11-08
8. [OrderOkr: OKR Goals Management: App Store](https://apps.apple.com/us/app/orderokr-okr-goals-management/id1609740590), Apple App Store, checked 2026-11-08