Objectives & Key Results (OKRs)

Avi Siegel

•

5 min read

What are OKRs?

OKRs stands for Objectives & Key Results. They are a framework that aligns organizations around ambitious goals and measurable outcomes. Proper OKRs force teams to focus on delivering meaningful value while pushing boundaries of what seems possible.

Why do companies set OKRs?

The primary goals of OKRs are to:

  • Align the entire organization on what matters for business success (team by team)

  • Encourage teams to think big and take calculated risks

  • Establish quantitative proof of success

  • Foster transparency and accountability

How do you set great OKRs?

As you could expect, there are two main components of setting OKRs - setting your Objectives, and setting your Key Results. It might sound simple, but doing OKRs well is easier said than done.

Define Objectives

Objectives are your goals. Not just any goals mind you - your BIG goals. They are the kind of goals that, if you can meet them, will almost assuredly mean success for your business, product, or team.

These goals should be:

  • Ambitious enough to drive real change

  • Focused on clear business success

  • Connect to company strategy (and in particular, the North Star)

Some would advise using "BHAG"s here - Big Hairy Audacious Goals. Yes, Objectives should indeed stretch the team, but they shouldn’t be so outlandish that they discourage effort.

  • "Big" - yes.

  • "Hairy" - huh?

  • "Audacious" - implies a level of foolishness and impossibility that is unwise

  • "Goal" - of course

We’re left with "Big Goals" - and that’s perfect for your Objectives.

Define Key Results

Key Results are the way you measure your progress against your goals. It’s how you prove that you’re doing work that actually matters, specifically as tied to your Objectives.

These metrics should be:

  • Specific: Clearly defined and unambiguous

  • Metric-driven: Based on quantitative measurements (including starting points)

  • Ambitious: Just out of comfortable reach to drive innovation

  • Relevant: Directly tied to the Objective

  • Value-based: Focused on outcomes, not outputs

Some would focus on the "SMART" acronym here. Note that what we have above is slightly different:

  • The M stands for Metric-driven, not Measurable - the KRs need to include a metric that matters to the business, not simply have a number attached to them

  • The A stands for Ambitious, not Achievable - the KRs should be a stretch for the team to push out-of-the-box thinking, not merely "doable"

  • The T was removed - it’s indeed important for the KRs to be Time-bound so they’re actually hit, but A) this is implicit in the fact that this is all typically done in quarterly or annual OKR setting, and B) this is often "impossible" due to how OKRs are typically managed, which is to say that results can’t be calculated at the end of the quarter or year because that’s the deadline for the work to be done, not the metric to be hit

  • The V was added - a focus on actual value is missing in SMART, and is arguably the most important requirement for defining good KRs (because if you’re not adding value, then why are you doing what you’re doing?)

We’re left with SMARV. Not quite as memorable, but certainly more representative of what you should be shooting for.

Pro Tip: Resist the urge to use output metrics like "launch feature X" or "fix Y bugs". Focus instead on the outcomes - the value - that those outputs should deliver, like "reduce error rate from 2% to 0.2%".

Who should set OKRs?

The classic answer is that OKRs cascade down from leadership to departments to teams. Yes, all those hierarchical goals should be tied together - but they shouldn’t be handed down from top to bottom as may be implied.

  • The North Star should come from the executive team in general, if not the CEO or founder(s) specifically

  • Layers of leadership and management should set OKRs as appropriate to funnel specific goals toward individual teams - note that this does not mean there is a level of OKRs for every level of management, but (especially for larger organizations) there will certainly be some breakdown here

  • Teams themselves (specifically, product managers owning respective capability areas) should drive the creation of their own OKRs (although the OKRs still need to be approved by higher levels in the chain)

The result here is a tree of OKRs, with each branch connected to its parent and children as appropriate.

Pro Tip: Every layer of OKRs (besides the North Star) should be a collaboration - a negotiation between higher level goals and lower level capabilities and responsibilities. Leadership’s role is to ensure alignment and push teams to stretch their abilities, whereas individual teams have the necessary context to set ambitious but achievable targets. If OKRs strictly go top-down or bottom-up, the results will likely be either unrealistic or not directed by the vision, respectively.

When should you measure OKRs?

There’s an awkward truth about OKR timing that nobody talks about: If you complete your work at the end of the quarter, that’s when you start measuring results, not when you finish measuring them. Instead, though, teams move on to setting the next quarter’s OKRs, and never look back.

OKRs should not drop out of view once a quarter is complete. Keep tracking metrics from previous OKRs until you have statistical significance as to whether the work drove the desired results.

Pro Tip: Create a "metrics review" process (or add it as part of the existing OKR process) to look back at previous quarters’ OKRs. Otherwise, you’ll never know if your work actually delivered the intended value.

Best practices surrounding OKRs

  • Focus on outcomes over outputs. "Launch feature X" isn’t a Key Result - the value that feature delivers is.

  • Use real metrics, not tasks. OKRs are not about tracking work - there are other tools for that.

  • Each team should have their own OKRs. If you can’t tell who owns an OKR, it’s probably too generic.

  • Don’t set impossible targets. Ambitious is good; impossible is demotivating. There’s a fine line here that’s tough to draw, so ensure that teams are deeply embedded into the process and properly empowered.

  • Be patient with improvement. OKRs are quite difficult to do well (see examples of bad OKRs here). Focus on iteratively getting better each quarter.

  • Measure what matters. Avoid vanity metrics and focus on real indicators of success.

  • Actually measure results (later). By definition, OKRs won’t be hit within the quarter they are assigned, because results can’t be measured the instant the work is completed. That means you have to come back later - quarter after quarter - to check back in on results. If you skip this part, you’ll never definitively know if your work made a difference.

  • Remember Goodhart’s Law: Be wary of metrics that can be gamed (i.e., all of them).