OKR (Objectives and Key Results)
1. Overview
A. Definition
OKR (Objectives and Key Results) is a goal-setting and performance management framework in which an organization sets ambitious Objectives and defines their degree of attainment as measurable Key Results, then executes, reviews, and aligns them on short cycles (typically quarterly). It was created by Intel's Andy Grove in the 1970s and spread to Google by John Doerr in 1999, becoming a standard Silicon Valley management tool.
The core insight of OKR is to 'separate and clarify what you want to achieve (Objective) from how you will know whether you achieved it (Key Results)'. Many organizations stop at slogans that have direction but cannot be measured, such as "improve the customer experience", or conversely sink into metrics that have numbers but no 'why', such as "reduce incidents by 10%". OKR pairs an inspirational qualitative objective (Why·What) with the quantitative results (How much) that prove its attainment, securing both direction and measurability at once.
Another core element is transparency and alignment. Whereas traditional goal management was a private contract known only to the supervisor and the individual, OKR publishes the goals of the whole company, teams, and individuals to everyone. Because anyone can see the CEO's OKRs and a neighboring team's OKRs, each person can judge for themselves how their work connects to the organizational strategy and collaborate horizontally. In other words, OKR is, before it is a performance measurement tool, a communication tool that translates strategy into an execution language and alignment for the whole organization.
B. Background and necessity
The dominant goal-management technique of the mid-20th century was Peter Drucker's MBO (Management by Objectives). MBO brought the innovation of management by objectives, but it gradually became rigid due to its long annual cycle of setting and evaluation, top-down allocation, and direct linkage to rewards. Goals were fixed for a whole year and could not respond to rapidly changing markets, and once rewards were at stake, members showed defensive behavior, proposing only 'safe goals' that were easy to achieve.
As digital transformation and the spread of agile increased the uncertainty and volatility of the business environment, year-long fixed plans became obsolete the moment they were drawn up. Organizations needed a new framework that re-adjusts direction on shorter cycles and motivates members to commit to ambitious goals on their own. OKR, proven by Intel and Google, met this need in that it ① realigns nimbly on a quarterly cycle, ② aligns the organization by publishing goals, and ③ encourages Stretch by separating itself from rewards. In short, the rise of OKR is a product of a management-paradigm shift: "make things measurable, but use measurement as a tool for execution and focus, not for evaluation".
2. Structure and components of OKR
The overall structure of OKR is a hierarchical, alignment-oriented form in which a higher organization's goals cascade downward, multiple Key Results hang off a single Objective, and Initiatives (execution tasks) to achieve each Key Result are connected. The overall structure diagram below shows both the alignment flowing from company → team → individual and the vertical connection of O–KR–Initiative.
flowchart TB
CO["Company Objective(company vision·strategy)"] --> CKR1["Company KR 1"]
CO --> CKR2["Company KR 2"]
CKR1 -. align .-> TO["Team Objective"]
CKR2 -. align .-> TO
TO --> TKR1["Team KR 1(quantitative metric)"]
TO --> TKR2["Team KR 2(quantitative metric)"]
TKR1 --> I1["Initiative(execution task)"]
TKR2 --> I2["Initiative(execution task)"]
TKR1 -. align .-> PO["Individual Objective"]
style CO fill:#e8f0fe,stroke:#2f6fed,stroke-width:2px
style TO fill:#fef3e8,stroke:#ed8f2f,stroke-width:1px
style PO fill:#eafaf0,stroke:#2faf5f,stroke-width:1px
The components are not a simple parallel list but a logical hierarchy of 'why – what and how much – how'. Let us examine each element together with its principle.
A. Objective — direction and inspiration
An Objective answers "where is the point we want to reach" and is a qualitative, aspirational goal. It is expressed not as a number but as a single sentence, and must have both the inspiration that makes members' hearts race (Inspirational) and the concreteness to create meaningful change within the quarter. For example, "create the most loved delivery experience in the domestic market" provides both direction and motivation.
The condition for a good Objective is focus. The number of Objectives an organization can handle in one cycle is usually limited to 3–5. If goals exceed ten, that is not a set of goals but a to-do list; resources are scattered and none are properly achieved. This is why OKR is called a 'tool of selection and focus'. In fact, Google bundles Objectives per organization and individual into a small number each quarter, forcing capability to be concentrated on the most important few.
A common failure in setting Objectives is putting a 'measurable number' in the Objective's place. "Achieve 10 billion in revenue" is closer to a Key Result than an Objective. An Objective must contain why that number matters and what change it aims for, so as to draw out members' intrinsic motivation and creative execution.
B. Key Results — measurement and evidence
A Key Result answers "how will you prove that you reached that objective" and is a quantitative, verifiable outcome metric. There are usually 2–5 KRs per Objective. A KR must be an Outcome, not an Activity. "Run 3 ad campaigns" is an activity, but "raise the new-signup conversion rate from 2.1% to 3.5%" is an outcome. Treating activities as KRs leads to mistaking a busy-but-resultless state for 'achievement'.
The core of a good KR is specifying a Baseline – target value – measurement method. Write not "improve NPS" but "raise NPS from 32 to 45 (based on the monthly survey)". Only then can an attainment rate of 0.0–1.0 be objectively calculated at cycle-end (e.g. 32→40 yields (40−32)/(45−32)≈0.62). A KR that cannot be measured or is ambiguous fails to create a meaningful conversation at mid-cycle check-ins.
Stretch is also an important property of a KR. Google uses a benchmark of '10-10-80' or '0.6–0.7 is ideal'. If all KRs are achieved 100%, it is interpreted as a signal that the goal was too easy. This design is possible precisely because OKR is separated from rewards, providing a psychological safety net so that members challenge ambitious goals without fearing failure.
C. Initiatives — the means
An Initiative is the activity·project actually carried out to move a KR. Managing it separately from the OKR body (O·KR) is key, because an Initiative is 'an input we can control' while a KR is 'the output we hope that input produces'. For example, to 'raise the conversion rate (KR)', Initiatives such as 'redesign the onboarding screen' and 'build an A/B testing system' are set.
Confusing Initiatives with KRs reduces OKR to a task-management tool. In practice it is recommended to link 2–4 core Initiatives per KR, and to adopt a dual structure that checks Initiative progress weekly and KR attainment quarterly. This lets you separately see "did we work hard (Initiative)" and "did the result change (KR)".
The table below compares and organizes the character and examples of the three components (an aid to the prose explanation).
| Component | Question it answers | Character | Example |
|---|---|---|---|
| Objective | Where are we going? | Qualitative·aspirational·direction | Deliver the best delivery experience in the country |
| Key Result | How do we know we arrived? | Quantitative·outcome·measured | Same-day delivery ratio 60%→85%, NPS 32→45 |
| Initiative | What will we do? | Activity·input·means | Open 2 regional logistics hubs, automate picking |
3. OKR operating process (cycle)
OKR is not a plan set once and done, but a rhythm (Cadence) in which set–align–execute/check-in–evaluate/retrospect cycles every quarter. The detailed process diagram below shows the structure in which the quarterly cycle and weekly checks overlap.
flowchart LR
A["1.Set(establish O·KR at quarter start)"] --> B["2.Align(publish·connect up·down·across)"]
B --> C["3.Execute·check-in(weekly CFR·progress review)"]
C --> D["4.Evaluate·grade(0.0~1.0 Grading at cycle end)"]
D --> E["5.Retrospective(retrospective·learning)"]
E --> A
C -.-> C
style A fill:#e8f0fe,stroke:#2f6fed,stroke-width:1px
style D fill:#fef3e8,stroke:#ed8f2f,stroke-width:1px
In the setting stage, each team sets its own OKRs based on the company OKRs derived from the higher strategy. If only full top-down cascading is used here, the rigidity of MBO is reproduced, so it is mixed with bottom-up proposals. Typically about 60% of company OKRs are designed so that teams propose them voluntarily, raising commitment.
In the alignment stage, the established OKRs are published across the company and inter-team dependencies are coordinated. For example, if 'the marketing team's lead-conversion KR' and 'the sales team's revenue KR' presuppose each other, the two teams form a horizontal contract while looking at the published OKRs. Transparency is the precondition of alignment.
The execute·check-in stage is the key that decides OKR success or failure. By running CFR (Conversation·Feedback·Recognition), emphasized by John Doerr, as a weekly rhythm, KR progress (On track/At risk/Off track) is shared as a traffic light and obstacles are removed early. Without check-ins, OKR becomes a decoration written down at the start of the quarter and forgotten.
In the evaluate·retrospect stage, each KR is graded 0.0–1.0, but the focus is on 'what was learned' rather than the score itself. Even a 0.4, if the insight gained in the process changes the next quarter's strategy, is a valuable failure. This retrospective feeds back into the next cycle's setting, accumulating organizational learning.
4. Core principles and types
The design principles underpinning OKR can be summarized in four. First, Focus — concentrate capability on a few important things. Second, Align — achieve vertical·horizontal alignment through publishing and bidirectional connection. Third, Track — re-adjust based on data through regular check-ins. Fourth, Stretch — pursue leap-forward results (10x) with ambitious goals beyond the comfort zone.
In terms of type, OKR is divided into Committed OKRs and Aspirational OKRs. A Committed OKR is an operational promise that must be 100% achieved (e.g. regulatory compliance, availability SLA), while an Aspirational OKR is a stretch goal for which 60–70% attainment is seen as ideal. The two types are run as a mix, but if the shortfall of an Aspirational OKR is punished as failure, the organization immediately turns conservative, so the distinction must be maintained. Also, by alignment method it is divided into top-down·bottom-up·hybrid, and a healthy organization chooses the hybrid.
5. Comparison: OKR vs MBO vs KPI vs BSC
These techniques all belong to the category of 'goal·performance management', but their aims differ. More important than knowing the differences is understanding why the differences arise and their practical implications.
| Category | OKR | MBO | KPI | BSC |
|---|---|---|---|---|
| Core purpose | Focus·align·stretch | Goal agreement·control | State monitoring | Balanced strategy measurement |
| Cycle | Quarterly (short) | Annual | Continuous | Annual·quarterly |
| Openness | Fully public | Private (1:1) | Within department | Management-centric |
| Reward linkage | Separation recommended | Direct linkage | Indirect linkage | Indirect linkage |
| Goal difficulty | Challenging (Stretch) | Achievable | Maintain·manage | Balance |
The biggest difference between OKR and MBO is separation from rewards and cycle. Because MBO directly linked goal achievement to evaluation·reward, the paradox arose of members choosing easy goals; OKR separated this and made challenge possible. The practical implication is clear — if you adopt OKR but tie it directly to the HR evaluation·bonus formula, it is OKR in name only and in effect regresses to MBO, and the challenge culture disappears.
The difference from KPI is 'direction vs state'. A KPI is a metric that continuously monitors the health state of an already-running process (e.g. server availability 99.9%), whereas OKR is a lever of change asking 'what will we change this quarter'. The two are complementary rather than opposed: keep what is to be stably managed as KPIs and handle what is to leap forward with OKR. BSC is a framework that measures strategy in balance across the four perspectives of financial·customer·process·learning, and OKR is a framework that picks 'the few to focus on now' among them and executes, so in practice a combination is also used in which BSC draws the strategy map and OKR unfolds the quarterly execution.
The practical implication of the selection criteria depends on the organization's situation and maturity. OKR, which encourages short cycles and challenge, suits growth-type·knowledge-intensive organizations where the environment changes fast and autonomy·engagement must be drawn out, while KPIs or Committed OKRs suit areas centered on operational metrics that must be achieved, such as regulatory compliance. Conversely, transplanting OKR as-is into a hierarchical·once-a-year evaluation culture easily regresses to MBO. Therefore, rather than "which technique is superior", a design that combines the three not exclusively but hierarchically — using the criterion of "which purpose (control·measurement·focus·challenge) to achieve at which cycle and openness level" — is closer to the correct answer from a professional-engineer perspective.
6. Application cases
Intel (the original case). In 1980 Andy Grove, under the Objective of 'recovering market dominance for the 8086 microprocessor', set concrete KRs such as benchmark performance and new design wins across the company and unfolded the 'Operation Crush' campaign as a quarterly OKR, reclaiming the market lost to Motorola. It is a representative case where OKR drew out company-wide focus in a crisis.
Google (the spreading case). Google still maintains the OKR that John Doerr spread in 1999 when it had about 40 employees, even after becoming an organization of over 20,000. It runs OKRs publicly, such as "improve search quality" and "reach 20 million monthly active users of Chrome", and institutionalized challenge by setting the ideal attainment rate of Aspirational OKRs at 0.6–0.7. Google's growth itself is cited as evidence of OKR's scalability (small scale→large scale).
Domestic application. Domestically too, many IT·platform companies and public institutions have adopted quarterly OKRs to connect company strategy with team execution. However, 'adoption failure' cases are also reported where it conflicts with the once-a-year HR evaluation·bonus system, or goals are merely declared and then neglected without a check-in (CFR) culture. The common feature of success cases is that they kept the three conditions: ① focus with a few Objectives, ② establish weekly check-ins, ③ separation from evaluation·rewards.
The lesson running through the cases is that OKR is not a 'goal-writing template' but 'the way the organization works'. Intel's crisis recovery, Google's expansion, and the domestic split between success and failure were all divided not by the sophistication of the template but by whether the operating principles of focus·transparency·check-in·separation were observed. Therefore adopting organizations must place greater weight on leaders' behavioral change and the embedding of organizational culture than on tool selection.
7. Deep dive: latest trends and expected exam directions
Recent OKR operation is evolving toward always-on visualization based on dedicated software. There is a clear flow of integrating OKRs, once managed in spreadsheets, into collaboration tools (e.g. work-management·performance-management SaaS) to automate real-time progress rates, alignment graphs, and check-in reminders. This mitigates the chronic failure of 'declare at quarter start, then neglect' through check-in automation.
Second, combination with agile·DevOps is spreading. Matching quarterly OKRs with sprint·release rhythms and connecting team performance to outcome-centric KRs such as DORA metrics (deployment frequency·change lead time, etc.) is increasing. This is because OKR's 'outcome-centric' philosophy meets agile's 'value-delivery' philosophy.
Third, integration with generative AI appears at an early stage. Uses are being tried where AI analyzes past data to propose KR drafts·baselines, or summarizes check-in comments to give leaders early warning of risk signals. However, in that the 'why' of a goal and strategic judgment remain the province of humans, AI settles in as an auxiliary means.
From a professional-engineer perspective, expected exam directions are likely to be essay questions asking about application strategy and trade-offs rather than simple definitions, such as ① comparison and selection criteria of OKR with MBO/KPI/BSC, ② ways to secure consistency with organizational culture·evaluation systems when adopting OKR, ③ OKR linkage design in organizations transforming to agile·DevOps, ④ analysis of success·failure factors of OKR landing in public·large-scale organizations.
8. Considerations and implications
To land OKR successfully, from a professional-engineer perspective the following must be considered comprehensively.
- Strategy of separation from evaluation·rewards: Tying OKR directly to the bonus formula collapses the challenge culture. The key is a 'separation design' that keeps a separate competency·contribution-centric evaluation system and runs OKR as a direction-alignment·learning tool. However, since full separation can weaken accountability, there is a trade-off of compromising by indirectly reflecting engagement·learning perspectives.
- Embedding the check-in (CFR) culture: OKR success or failure depends on the continuity of weekly check-ins more than on the quality of goal writing. If leaders do not run 1:1 conversation·feedback·recognition as a rhythm, OKR degenerates into a quarter-start decoration. Leadership·organizational-culture change must precede tool adoption.
- Gradual adoption matched to organizational maturity: It is safer to pilot in a trial organization for a quarter or two before scaling, rather than a company-wide bulk adoption. In particular, public·large enterprises accustomed to hierarchical·once-a-year evaluation must resolve conflicts with the existing performance-management system at the design stage.
- KR quality management and measurement system: An education·review system must be in place so that KRs are written as outcomes, not activities, specifying baseline·target value·measurement method. A KR for which data cannot be collected is empty, so securing a measurable data pipeline (analytics·BI) is a prerequisite.
- Outlook and related technologies: OKR will evolve toward 'always-on·automated alignment', combining with agile·DevOps·data-driven management (Data-Driven), performance-management SaaS, and generative-AI assistance. The professional engineer is required to view OKR not as a single technique but as the execution layer of digital-transformation governance, designing it as a three-tier structure of strategy (BSC) – execution (OKR) – operation (KPI).
References
- John Doerr, "Measure What Matters" (2018), https://www.whatmatters.com/
- Google re:Work, "Guide: Set goals with OKRs", https://rework.withgoogle.com/guides/set-goals-with-okrs/
- Rick Klau (Google Ventures), "How Google sets goals: OKRs", https://library.gv.com/how-google-sets-goals-okrs-a1f69b0b72c7
In one line: OKR is a goal-setting·execution framework that pairs an inspirational qualitative Objective with measurable Key Results, publishes and aligns them, and draws out focus·challenge·learning through reward-separated quarterly-cycle check-ins (CFR).