Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Comparisons

OKRs vs KPIs

Objectives and key results compared with key performance indicators: committing to change against monitoring health, and how marketing teams connect the two.
Comparisonbeginner8 min readUpdated

OKRs and KPIs at a glance

Key performance indicators are the measures that prove whether objectives are being met. The guide to setting objectives and KPIs builds them as a chain from the business goal down: a marketing objective that would move the goal, one primary KPI that proves it, a target and a date, and diagnostic metrics that explain movement in the KPI. Avinash Kaushik's Digital Marketing and Measurement Model, a widely cited version of the same idea, starts from business objectives and works down through goals, KPIs, targets and the segments to analyse. A KPI set is meant to be stable. It describes the health of the engine, it is reported every week or month, and it changes only when the business or its strategy changes.

OKRs, objectives and key results, are a goal-setting method rather than a type of metric. The north star metric and OKRs framework covers how to write them, how to score them and how they connect to a north star. Google's re:Work guide to OKRs, cited there, describes objectives as ambitious and concrete and key results as measurable and verifiable, kept few in number and visible across the organisation. OKRs are meant to change. Each cycle the team picks a small number of things it intends to improve, commits to them, grades the result and sets new ones.

The difference is one of purpose, not of measurement. A KPI is a gauge on the dashboard; an OKR is a decision about which gauge to move and how far. Confusing them produces two familiar failures. A team that turns every KPI into a key result ends up with dozens of objectives and focus on none of them. A team that writes key results with no link to its KPIs does work that never shows up in the numbers leadership reads.

Compare table

OKRs vs KPIs

Row labelOKRsKPIs
PurposeCommit a team to a few changesMonitor the health of the engine
Time frameA cycle, usually a quarter, then resetContinuous; stable from year to year unless strategy changes
Number in useA small handful of objectives, each with a few key resultsA short list of primary measures, with diagnostics beneath them
AmbitionOften a stretch; partial achievement is normal for aspirational resultsTargets set at a level the plan depends on
OwnershipThe team that commits to the changeThe person or team responsible for the area the KPI measures
Review cadenceWeekly check-ins, a mid-cycle review and an end-of-cycle gradeWeekly or monthly reporting, with alerts when a KPI moves sharply
Link to pay or performanceBest kept separate, so key results stay ambitiousOften part of role expectations and reviews
Typical failureKey results written as tasks, or too many objectivesVanity metrics that move without changing the business

The last row is the useful one for a quick diagnosis. OKRs usually fail by drifting into activity, "launch the referral programme" instead of a result the launch should produce. KPIs usually fail by measuring something easy rather than something that matters, the Vanity Metrics that rise while revenue does not.

When to choose OKRs

Use OKRs when a team needs to change something, not just keep it running. A new channel to establish, a conversion problem to fix, a stage of the Funnel that loses too many people, a retention programme to build: each is a piece of work with a beginning and an end, and each benefits from a stated objective, a few measurable key results and a date.

OKRs are especially useful when there are more good ideas than capacity. Writing a small number of objectives forces the conversation about what will not be done this quarter, and making them visible across teams shows where priorities collide. The quarterly planning session tutorial walks through drafting objectives and key results, scoring initiatives and checking the list against capacity in one session.

OKRs are less useful for work that does not change from quarter to quarter. Keeping the email programme running, publishing on schedule and paying suppliers on time are responsibilities, and writing them as objectives clutters the list and dilutes the ones that matter. Those belong in the KPI set and in role expectations.

When to choose KPIs

Use KPIs to know whether the business is healthy at every moment, including in areas nobody is actively working on. The KPI set should cover the whole engine: acquisition, conversion, retention and economics, each with a primary measure and a target. Because it is stable, it builds a history that makes seasonal patterns and slow declines visible, which a set of objectives that changes every quarter cannot do.

KPIs are also the foundation for good OKRs. A team that does not know its baseline cannot write a meaningful key result, and a team without a clear view of how its metrics relate cannot choose the right one to move. A KPI tree breaks one outcome into the metrics that mathematically produce it, which shows which branch is weakest and therefore which KPI deserves an objective this quarter.

Where KPIs fall short is direction. A dashboard shows that conversion is flat or that cost per lead has risen; it does not say what the team will do about it or by when. Leaving that implicit is how a known problem survives several quarters of reporting.

Illustrative example — not a real result

A language-learning app turns a KPI into a key result

Wrenhallow, a language-learning app, reports four primary KPIs every month: new trial starts, trial-to-paid conversion, monthly churn and cost per paying subscriber. Trial-to-paid conversion has sat at 11% for three quarters while the other KPIs are on target. The KPI tree shows that most trials end without the learner finishing a first lesson. For the next quarter the team writes one objective, "make the first week of a trial impossible to waste", with three key results: first-lesson completion within two days rises from 38% to 55%, trial-to-paid conversion rises from 11% to 14%, and support contacts per trial do not rise. The KPIs stay on the monthly report unchanged; the OKR is graded at the end of the quarter and replaced.

Decision table

SituationUseWhy
Reporting the health of marketing to leadership each monthKPIsStable measures build a history and cover the whole engine
Fixing a weak stage of the funnel this quarterOKRsA stated change with measurable results and a date
Too many initiatives for the team's capacityOKRsA small set of objectives forces the trade-offs into the open
Ongoing responsibilities that do not changeKPIs and role expectationsWriting them as objectives dilutes the real priorities
Choosing which metric deserves effort nextKPIs, read through a KPI treeThe weakest branch points to the right objective
Setting a stretch goal for a new capabilityOKRs, marked aspirationalPartial achievement is expected and still useful
Performance reviews for individualsKPIs and committed goalsTying stretch key results to pay makes them safe

Turning a KPI into a key result

The simplest link between the two systems is to take a KPI that needs to move and write it as a key result. Three steps keep it honest. First, record the baseline from the KPI report at the moment the key result is written, so the end-of-cycle grade is a reading, not a negotiation. Second, set the new level with the cycle in mind: a KPI that moves slowly, such as retention over a long customer life, may need a leading input as its key result, with the KPI itself watched as confirmation. Third, add a guardrail where moving the KPI could damage something else, such as a lead-volume key result paired with a lead-quality one, because whatever is made a target will be pushed by any means available. The same split holds when a team's plan is broken down person by person: the tutorial on setting objectives for each marketer leaves ongoing responsibilities as KPIs and gives each person outcome objectives traced to a team key result.

Not every key result needs to be an existing KPI. A team building something new, such as a measurement capability or a channel, may write key results that describe outcomes on the way, like the share of spend covered by a lift test. Those should still be outcomes rather than tasks, and when the work is done the new capability often produces a KPI of its own.

The link also runs the other way. When an objective succeeds, the new level becomes the KPI target for the following quarters and the OKR is retired. When the same KPI keeps reappearing as a key result, cycle after cycle, the problem is probably structural, and the KPI tree framework is the place to look for the branch that is really holding it back. For a sense of how the same choice between motions plays out across a company, the comparison of product-led and sales-led growth shows how the primary metrics differ by motion.

Common mistakes

  • Turning every KPI into an OKR. A dozen objectives is a dashboard with a new heading, and focus disappears.
  • Writing key results as tasks. "Publish ten articles" is activity; the key result is what the articles should change.
  • Setting key results with no baseline. Without the starting value, the grade becomes an argument.
  • Choosing KPIs because they are easy to measure. Followers, impressions and page views rise without moving the business unless they are linked to an outcome.
  • Linking stretch key results to pay. People then write key results they are certain to hit.
  • Changing KPIs every quarter. The history that makes them useful is lost, and declines go unnoticed.
  • Leaving no owner for a KPI nobody is working on. Health measures still need someone who notices when they slip.

Frequently asked questions

Is a key result the same as a KPI?

Often it is the same metric with a different job. A KPI is watched continuously to show health; a key result commits a team to moving a metric to a new level within a cycle. When the key result is a KPI pushed to a new level, the two are connected and the team's work shows up in the numbers leadership already reads. A key result can also be a new measure, such as a milestone on the way to a capability the team does not yet have, but it should still be an outcome rather than a task.

Should a marketing team have both OKRs and KPIs?

Most benefit from both. KPIs cover the whole engine all the time, so problems are noticed wherever they appear. OKRs focus the team on the few changes that matter most this quarter, which KPIs alone do not do because they say where things are, not where effort should go. A team with only KPIs tends to defend the status quo; a team with only OKRs can miss a quiet decline in something nobody is working on.

Should OKRs be linked to pay?

Google's re:Work guide treats OKR grades as a tool for learning, and the reasoning is practical: when pay depends on the grade, people write key results they are sure to hit, and those teach nothing. KPIs are more often part of performance reviews, because they describe ongoing responsibilities. If any goal does feed into appraisal, label it as committed and keep stretch key results out of it.

Terms used on this page

Frameworks

Where this page is used

Also related

Sources

  1. Guide: Set goals with OKRs — Google re:Work
  2. Digital Marketing and Measurement Model — Avinash Kaushik

Related