Yaniv Preiss

We’re going deep into OKRs, and specifically, about mistakes that make them ineffective or less effective. 

Many companies I meet use OKRs, but only a handful use them well.

I am going to list 41 mistakes, and the purpose is not to remember all of them, but to be aware of the nuances and examine them when you define yours.

Once you’re aware, noticing mistakes becomes easier, but still, making great OKRs is a challenge. For the first few times, you will probably make mistakes and continue to improve with each time. Having mistakes in OKRs is not the end of the world; it’s an opportunity to do better next time. Don’t let perfect be the enemy of good.

1. Setting tasks and initiatives instead of outcomes
A classic trap. Teams write Key Results like:

• Launch a new website
• Move all customers to the new UI
• Talk to 5 customers
• Migrate to version 3
• Upgrade Kubernetes
• Salaries paid on time

But these are tasks, not outcome. Projects describe what you’ll do. Key Results measure what you will have achieved.

Maximum churn of 2% is a Key Result.

2. Vague objectives
Objectives like “Improve communication”, “Make progress with customers” or “Be more innovative” sound nice but mean nothing. They are gut feeling, subjective and interpretable, and one can always argue they’ve been achieved.

A good Objective should inspire and guide. It should tell you what success looks like without needing a paragraph of explanation, for example: “Be the number one choice for small businesses in Germany.”

3. Making Key Results unmeasurable
If you can’t measure it, you can’t manage it.

“Increase customer satisfaction”, “improve system security” – by how much? What’s the current value?

Add numbers and clarity: “Customer satisfaction score of at least 7.5.”

4. Having too many OKRs.
I recently saw a team of 7 members that had 5 Objectives and 6-8 Key results in each. Some teams set so many, thinking they’re being thorough. In reality, they’re diluting focus. Who can work on, track and make real progress on 35 key results?

1-2 solid Objectives per team per quarter, with 2-3 KR each are usually plenty. Less is more. Focus helps with business impact.

5. X Items done
When it comes to measuring, some mistakenly measure tasks instead of outcome. They may sound like OKRs, like “4 initiatives done”, not even stating what these initiatives should address.

“4 letters of intent received” is a key result.

6. Copying last quarter’s OKRs
Continuing the same goals is totally fine, but if you constantly roll them over, you’re not learning or improving. Each cycle should reflect what’s changed: the market, the strategy, or your priorities.

7. Creating “business as usual” OKRs
OKRs are meant for change, not for routine operations. If it’s part of someone’s daily job, it doesn’t need to be an OKR.
OKRs should stretch the team toward improvement or transformation.
You don’t need OKRs for “Maintain uptime at 99%” when this is your current baseline.

“Reach uptime of 99.94%” could be a key result.

8. Setting OKRs without strategy
OKRs aren’t a substitute for strategy. They should translate your strategy.
If you can’t link each OKR to a strategic priority, stop and clarify your direction first.

Every Objective should connect to purpose – why this goal matters, who it impacts, and what success will change.

9. Writing top-down OKRs only, without involving the team
People commit to what they help create (IKEA effect).

When OKRs are always imposed on teams, motivation drops. Involve your team early in definition, not just execution.

Top-down alignment is good, but bottom-up input builds ownership. Let teams shape their own contributions to the bigger goals. They know their domain best.

A mix of top-down and bottom-up is also advisable.

10. Defining OKRs for individuals
OKRs are about collaboration, not isolation.

When OKRs are set by managers to individuals, they optimize locally, often at the expense of team outcomes. If you had a personal one and a team one at conflict, which one would you pursue?

Keep them at team or department level.

11. Linking OKRs to bonuses or performance reviews
The moment OKRs affect pay, people start sandbagging. They’ll pick easy goals, play it safe, and stop being ambitious. They will also pursue them even if they become irrelevant. OKRs should inspire learning, not fear.

12. Treating OKRs as a to-do list
If your OKRs look like a list of tasks, you’re managing activity, not impact.
OKRs are not a project tracker, they’re a focus and alignment mechanism.

13. Not checking progress regularly
“Set and forget” OKRs die quietly.

You need to revisit them weekly or biweekly. Check what’s moving, what’s stuck, and what needs changing. If there is no progress for 11 weeks, there is a very slight chance that it will move in the last week of the quarter.

If you wait until the quarter ends to check progress, you’re too late to consider other options, make decisions and improve.

Regular check-ins turn OKRs from a report into a living process.

14. Ignoring human confidence
Tracking numbers is a must, and additionally checking confidence in those numbers is even better.

There may be information that the team knows and is not evident in data yet. If the team’s confidence drops, that’s a red flag – act on it early.

15. Writing binary Key Results
“Client X signed” gives you a yes/no result.
Use incremental measurements instead of all-or-nothing, like “30 interested customers signed up for beta.”

16. Keeping OKRs private
Hidden OKRs kill alignment.

Make them visible across teams and departments so everyone understands how their work connects. Transparency drives collaboration.

17. Letting every team use a different format
If marketing uses Excel, engineering uses a whiteboard, and sales uses a slide deck, you’ve lost cohesion.

Pick one format, one system, one cadence. This will help seeing the big picture and how all contribute to the company OKRs.

When all are in a single tool or document, you can easily see the combined progress.

18. Using only lagging indicators
Lagging indicators show what happened after the fact, like how many sales were done.

Balance them with leading indicators – early signals that tell you if you’re on track, like engagement rates or response rates to cold emails.

19. Forgetting dependencies
OKRs don’t exist in isolation.

If your key result depends on another team’s work, acknowledge it and sync early. Otherwise, you will need to convince and coordinate during the quarter and you might not get that cooperation, unable to achieve your OKRs.

20. Overcomplicating the scoring
Some teams build spreadsheets with weighted averages, complicated formulas and decimal points that take a long time to calculate.

Don’t. Keep it simple. The goal is learning, not accounting. You can use a proxy (something that represents the real value) and drop precision.

21. Ignoring ambition
If you hit 100% of your OKRs every time, you’re not stretching enough.

If you have an ambitious goal and hit 70%, that’s already a big accomplishment comparing to where you started.

22. Using OKRs as a communication substitute
Posting OKRs in a shared document doesn’t replace conversations between teams, stakeholders and management.

They should be a reason to talk, not a replacement for talking.

23. Writing OKRs once and never considering refining them
Markets may shift, priorities evolve, customers change.

Revisit your OKRs as you learn more, adjust targets, drop irrelevant ones, and celebrate small wins.

24. Not aligning across levels
Company, department, and team OKRs should connect like gears in a machine.

If each team moves in a different direction, you’ll waste effort instead of compounding it. And these do happen. I once witnessed a company that had two big modules and ran two opposing initiatives simultaneously – one to unify them, the other to extract a third one.

25. Limiting wording
Ground-struck OKRs, using the words “improve”, “increase”, “keep”, “reduce”, “maintain” with past references (from X to Y) – these limit the imagination, a mindset of doing a bit more of what we did till now. They anchor, or put emphasis on the current situation, assuming the current situation is god given, like in “Reduce checkout time from 2:50 to 2:30”.

A stronger one would be “Average checkout time of 2:30”.

26. Insignificance
OKRs that promote progress but don’t move the needle for the business.

“Improve sales from the previous quarter (from 100 to 101)”

27. Not having OKRs because you cannot measure
A measurement is only an observation to reduce uncertainty. Find a proxy and measure it.

For example, for the kindness of the security staff, sample the number of smiles at visitors and the complaints.

28. Confusing with KPIs
KPIs measure the performance of processes and activities already in place. OKRs are a change for the future. OKRs may be used to improve existing KPIs.

For example, “maintain eNPS of 4.1” is not an OKR because it’s keeping the current KPI. “eNPS of 5” when the current one is 4.1, is an OKR.

29. And
The word “and” means two different things, so they should be separate OKRs. Otherwise, if only one has been achieved, we still fail the OKR.

It also makes it hard to prioritize and make decisions, like in “satisfy customer A AND customer B”.
Some may be even conflicting: “Great company for investors AND employees”.

30. Initiatives that drive OKRs
This happens when the team gets tempted to think of what they want to do instead of the strategy, desired outcome, and how to move the needle.

31. Lack of data
“Improve number of support tickets from X to Y TBD”, or “reduce cost by X% TBD”

We didn’t care enough to measure baseline and/or know the outcome we are after – so why is it important to dedicate the whole quarter?

If the only argument is “Doing it just makes sense”, compare it to other things that make sense.

32. Percentages
Not inherently bad, just more confusing than absolute numbers.

If we have a conversion rate of 5% and we want to reach 10%, is it a 5% increase or a 100%?

Of course, sometimes we don’t have absolute numbers, like the number of errors in our API responses or the system uptime.

33. Negative formulation
Humans tend to pursue achievements rather than preventions.

Consider: “Up to 2% API error responses” vs. “Min 98% API success responses”, both sound ok-ish, but “Churn of up to 2 customers” vs. “Retain 42 customers” – prefer the positive.

34. Extreme dependency
Very high risk on a single source of impact: think of “Investor X gave LOI” – what if they cannot or will not, due to internal reasons you have zero impact on? Even if you learn about this at the beginning of the quarter, there is nothing you can do about it.

Better: “3 investors in the security industry gave LOIs”.

35. No balancing key restuls
With the excitement to hit a goal, we might unnoticeably damage another part of the business.

For example, we increased in-store revenue and inadvertently reduced online sales, so overall it was a loss for the company. It’s best practice to include a counter-measure or health metric for your KRs to make sure other important aspects are not negatively impacted.

36. Outcome likely to happen regardless
“Unblock launch of customer A” – while we have no evidence we need to do anything about it.

“Smooth holiday season” is acceptable when we know we need to deliberately act.

37. Weak words
“Can”, “enable”, “allow” are not really about impact.

For example: “3 teams can use the new dashboard”. And if they can but do not – what is the value? Is it a success?

Better: “3 teams use the new dashboard daily”.

38. Acronyms, Jargon, implicitness
“QBR of 800m for CLV of 5K via CR of 10%” – remember OKRs is for the whole organization and some stakeholders will need to understand “translate” each time.

39. Only OKR work
OKRs are about strategy. Business as usual will still be there. Products end-of-life, legal deadlines, engineering initiatives, tech debt, support, helping other teams, performance reviews…

You don’t need to capture every piece of work in OKRs, only the big rocks you move for the business.

40. Oscillating values
OKRs that use numbers that can change with every measurement, like the weekly number of support tickets: 12, then 40, then 6, then 20… are not the end of the world, but it makes it harder to track and can demotivate. They are also too easy to find excuses for missing.

Use average or cumulative: “Weekly average of 10” or “120 tickets cumulative”.

41. Not defining terms
Clarity is important.

“3 Enterprise customers booked 20 seats each” sounds great, but “Enterprise” is not defined. This can be a star below to keep it clean:

“3 Enterprise* customers booked 20 seats each”
Enterprise = more than $500M revenue

Or flattened: “3 customers with more than $500M revenue booked 20 seats each”.

---

There you have it – many OKRs mistakes. Unfortunately, many managers and teams continue to argue that such flawed OKRs are completely fine. Maybe, but you can be more effective!