People Analytics

As AI and People Analytics spread, the discussion around performance management is also changing. The old way of having people write goals manually and compile achievement rates by hand at the end of the quarter is gradually losing its persuasive power. When collaboration tools, work records, customer data, and HR data are connected, the progress of goals can become visible more often, in greater detail, and more automatically.

But this does not mean the automation of OKRs. AI can suggest goal statements, and People Analytics can quickly show changes in metrics. However, what should be regarded as an important goal, which indicators should be accepted as evidence of performance, and how to interpret the causes of underachievement remain matters of organizational judgment. In the age of AI, OKRs are moving in a direction where the operating system for interpretation and accountability becomes more important than the technique of writing goals.

AI cannot decide OKRs for an organization, but it lowers the cost of collecting evidence

Google’s OKR Playbook explains that Key Results should describe outcomes, not activities, and that evidence of completion should be available, credible, and easily discoverable. It gives examples such as documents, notes, and published metrics reports. This principle becomes even more important in the age of AI and People Analytics.

AI can identify signals of goal progress from meeting minutes, project management tools, customer feedback, and work documents. Delay signals that previously could be known only when a leader asked directly can now appear earlier through data. People Analytics can show indicators related to turnover, engagement, collaboration, capabilities, and productivity at the organizational-unit level. The cost of collecting evidence goes down.

However, having more evidence is different from having better goals. Organizations must distinguish whether the signals found by AI represent actual performance or simply the volume of activity. The number of documents written, meeting attendance, and tickets processed are easy to measure. But improvements in customer experience, strategy execution, and the accumulation of organizational capabilities require more careful interpretation. AI can gather evidence, but people must review what that evidence means.

The stronger People Analytics becomes, the stricter Key Results become

People Analytics is a powerful foundation that enables HR to make decisions based on data rather than intuition. AIHR describes People Analytics as a data-driven HR capability and presents types of analytics such as descriptive, diagnostic, predictive, and prescriptive analytics. As this trend grows stronger, OKR Key Results become stricter.

For example, under an Objective such as “strengthen leadership training,” if “conduct five training sessions” is set as a Key Result, AI and data tools can easily track that activity. But it is still an activity metric. As in the principle of Google’s playbook, Key Results should be outcomes, not activities. Organizations need to consider metrics closer to outcomes, such as “the rate at which new leaders hold one-on-one meetings with team members within 60 days,” “retention rate of key talent,” “project decision-making lead time,” and “recurrence rate of customer complaints.”

When there is more data, ambiguous goals are exposed more quickly. OKRs that have not defined what to measure cannot be placed on a dashboard. Conversely, if an organization sets goals only around what is easy to measure, it may miss important changes. Designing Key Results in the age of People Analytics is not about attaching numbers. It is about translating the outcomes the organization truly wants to change into the language of data.

Monthly check-ins become interpretation meetings, not data dashboards

Atlassian recommends that OKR operations score, analyze, and summarize every month. When AI and People Analytics are combined, monthly check-ins have more data. Progress rates, workloads, collaboration networks, employee experience, customer responses, and issue-delay signals can all be gathered on one screen.

Even so, check-ins should not end as dashboard reviews. CIPD explains that effective HR decision-making should be based on a combination of the best available evidence and critical thinking. Evidence helps judgment, but it does not replace judgment. What data tells us is closer to “what happened.” What leaders and HR need to ask is “why did this happen, and what will we change now?”

Therefore, OKR check-ins in the age of AI should become meetings for interpretation, not meetings for reading numbers. If an indicator has worsened, the meeting should focus on sharing causes rather than interrogating the person in charge. It must distinguish whether the goal was designed poorly, whether resources were insufficient, whether interdepartmental dependencies remained unresolved, or whether market conditions changed. Data is the starting point of the meeting, not its conclusion.

As data increases, HR’s question shifts from performance to accountability

As AI and People Analytics spread, HR can see more performance signals. But as signals increase, new risks also emerge. These include mistaking the volume of individual activity for performance, treating only measurable indicators as important, or connecting data to evaluation and rewards when data quality is low.

Google’s OKR Playbook explains that well-run OKRs clarify what is important, what should be optimized, and what tradeoffs should be made. This principle applies equally in the data era. HR’s question must not stop at “who is performing well?” It must shift to “what are we optimizing for which goal?”, “if we raise this metric, will other important values be damaged?”, and “is this result the outcome of individual effort, or of the system and resource allocation?”

In Korean companies in particular, data can quickly be linked to evaluation and compensation. That is why HR must first define the boundaries of data use. OKR progress data can serve as reference material for performance conversations, but it is risky for it to become an evaluation formula as is. AI-generated summaries should also be presented with reviewable evidence. The more data there is, the more responsible rules of interpretation are needed.

In the age of AI, OKRs need operating governance more than automation tools

The AI Index states that its purpose is to track, collect, organize, and visualize AI-related data to help policymakers, researchers, business leaders, and the public understand AI. This trend also has implications for HR. Performance management will handle more data going forward. But the more data there is, the more organizations must decide what to measure, who can access it, and what decisions it will be used for.

OKR operations in the age of AI require at least three types of governance. First is a quality standard for goal data. Organizations must decide which indicators will be accepted as Key Results and which data will be used only as reference. Second is a standard for interpretation authority. Organizations must decide who will interpret AI summaries, dashboards, and People Analytics results, and in which meetings those interpretations will be finalized. Third is a standard for connection to evaluation and rewards. Organizations must separate how far OKR data serves as evidence for performance conversations and from what point it becomes material for compensation judgments.

For example, it is risky to interpret the number of messages in collaboration tools or meeting attendance rates directly as “engagement” or “collaboration performance.” Conversely, indicators connected to work outcomes, such as customer response time, reduction in recurring complaints, and decision-making lead time for key projects, can be good starting points for OKR interpretation. HR must consider not the convenience of indicators, but their relevance to outcomes, privacy and labor risks, and explainability to employees.

The starting point of this OKR series was the recognition that OKRs are not a goal-management template but an operating system for performance management. In the age of AI and People Analytics, this perspective becomes even more important. AI can create goal statements more quickly. Systems can show progress rates more often. But what the organization chooses, what it gives up, and what evidence it recognizes as performance are not automated.

Ultimately, OKRs in the age of AI are not a matter of technology adoption. They are a matter of organizational operations that responsibly interpret more data. HR’s role also shifts from tool administrator to designer of the language of performance. We are not entering an era in which AI manages goals on behalf of organizations, but an era in which organizations must make more sophisticated judgments around the evidence that AI reveals.

OKR is no longer an unfamiliar term in Korean companies. Many organizations have already conducted OKR training, created quarterly goal templates, and some have even tried placing OKRs inside their performance management systems. Yet as implementation experience accumulates, the same question keeps coming back. Why do OKRs feel new at first, only to become similar to existing goal management after a few months?

The answer to this question lies not in the tool, but in the organization’s operating language. In Korean companies, OKRs collide at the same time with evaluation memory, reporting culture, interdepartmental accountability structures, and leaders’ decision-making styles. Introducing a template is not enough. For OKRs to take root, the way goals are interpreted and adjusted must change before the way goals are written.

In Korean companies, OKRs collide with evaluation memory before they collide with systems

What Matters explains, in comparing OKRs with MBO, that OKRs spread as a quarterly practice with a philosophy separated from compensation. This explanation is especially important for Korean companies. In many organizations, goals are remembered as evaluation forms. The experience is strong: goals are set at the beginning of the year, achievement rates are checked at year-end, and the results are connected to ratings and rewards.

When OKRs are introduced while this memory remains, employees naturally behave defensively. Even if they are told to write ambitious goals, they choose safe goals if they feel those goals could hurt them in evaluation. Even if they are told to make goals public, they become cautious in wording if they think records of non-achievement will remain. Even if they are told to write collaborative goals, they avoid commitments that may disadvantage their own department if accountability is unclear.

Therefore, OKR adoption in Korean companies must begin not with the question, “Will we connect OKRs to evaluation?” but with explaining “What kind of language makes OKRs different from an evaluation form?” OKRs are not a system that eliminates evaluation. But it must be made clear that they are an operating language for adjusting priorities and execution direction during the quarter.

The first condition for adoption is not the number of goals, but agreement on what to give up

The Google OKR playbook explains that well-run OKRs make clear what is important, what should be optimized, and what tradeoffs should be made. Atlassian also suggests setting 1–3 Objectives and 3–5 Key Results per Objective. The message more important than the numbers is the limitation of priorities.

The moment OKRs in Korean companies return to conventional goal management is when the number of goals increases. If headquarters goals, team goals, individual goals, and project goals are all attached under the name OKR, OKR becomes a work list rather than a tool for focus. If leaders add only new goals without reducing existing work, employees receive OKRs as just another reporting item.

If an organization wants OKRs to take root, there must be agenda items that are explicitly removed from OKR meetings. It must decide what will not be done this quarter, what will be deferred, what will be managed only at a maintenance level, and what will be merged with another team’s work. It should become not an organization that uses OKRs, but an organization that reduces work because of OKRs. Only then will employees believe that the system actually changes priorities.

A particularly necessary device in Korean companies is a “stop list.” When a division head approves quarterly OKRs, they should not approve only new goals; they should also confirm which reports will be stopped, which meetings will be reduced, and which projects will be pushed to the next quarter. Without this list, frontline teams receive OKRs not as new priorities but as additional tasks layered on top of existing work.

The stronger the reporting culture, the more check-ins must become decision-making meetings

Atlassian explains that OKRs are set annually, refreshed quarterly, and progress is tracked monthly. Regular review is central to OKR adoption. However, in Korean companies with a strong reporting culture, check-ins easily become reporting meetings. The person in charge states the progress rate, the leader asks why things are delayed, and the minutes record “continue execution.”

OKRs are difficult to embed through this approach. OKR check-ins must be decision-making meetings, not reporting sessions. If progress is low, the question should not be who needs to try harder, but what needs to be adjusted. The meeting should decide whether to change priorities, reinforce resources, adjust the schedule of a dependent team, or revise the goal itself.

The Google playbook explains that when it appears difficult to achieve a committed OKR, escalation should happen immediately. This is closer to a conflict resolution process than a failure report. In Korean companies as well, OKR check-ins should be designed less as a place to report upward and more as a place to resolve conflicts with adjacent departments and for leaders to make choices.

Cross-department collaboration must be embedded in each department’s OKRs, not left as a slogan

The Google playbook explains that in cross-team OKRs, all groups that must actually participate should be included, and each group’s contribution should be specified in that group’s OKRs. This principle is especially important in organizations with strong boundaries between departments.

Korean companies emphasize collaboration, but the line of responsibility for collaborative goals can easily become blurred. A goal such as “improve customer experience” may connect marketing, sales, product, customer support, and HR. But if each department’s OKRs do not include its own contribution, deadline, and success criteria, the shared goal ends as a declaration. Collaboration is a positive word, but execution is weak when accountability is not specified.

When designing cross-team OKRs, HR should look not only at one shared goal but also at each department’s OKRs together. It should confirm which department provides data, which department changes the customer touchpoint, and which department adjusts operating policies. Even when departments look at the same goal, collaboration works only when the results each department will own are embedded in the document.

Korean-style OKR adoption is not localization, but translation of principles

The phrase “creating OKRs that fit Korean companies” is often used to mean weakening the system. The scope of disclosure is reduced, OKRs are connected slightly to evaluation, and Objective and Key Result fields are added to the existing KPI form. But this is less localization than a way of absorbing OKRs into the language of the existing system.

What is needed for adoption is the translation of principles. The principle that Key Results should be outcomes, not activities, remains valid in Korean companies. The principle that committed OKRs and aspirational OKRs should be distinguished also remains valid. The principle that cross-team OKRs should include the responsibilities of the groups that actually participate also remains valid. However, these principles must be explained and trained in relation to Korean companies’ evaluation systems, leadership reporting structures, and interdepartmental decision-making structures.

OKRs fail in Korean companies if they are imported unchanged, and they also fail if they are converted into conventional goal management. What is needed is not the translation of a template, but the translation of an operating language. It means asking “What should be adjusted?” instead of “Why did you miss it?”, asking “Which department’s contribution is missing from the document?” instead of “Who will be responsible?”, and asking “Is this goal a commitment or an aspiration?” instead of “What score is the achievement rate?”

When OKRs take root in Korean companies, it does not mean a foreign-style system has been introduced. It means the conversation around goals has changed. When leaders narrow priorities, HR clarifies the boundary between evaluation and operations, and departments specify accountability for shared goals, OKRs become not a system but a way of working.

In organizations that adopt OKRs, leaders often say, “Write the goals more clearly.” But the point where OKRs actually become unstable is not the wording but the operating process. Employees can write Objectives and turn Key Results into numbers. The problem arises when it is not decided what to give up for those goals, who will coordinate when conflicts occur, and what decisions will be made when progress deteriorates.

In OKRs, the leader’s role is not to be a cheerleader. The leader is the person who narrows priorities, coordinates resource conflicts, surfaces dependencies between teams, and turns check-in meetings into forums for decision-making. Without this role, OKRs become a document that asks employees for more goals.

What leaders must decide first is not goals but “work not to do”

The Google OKR playbook explains that well-run OKRs make clear to teams what is important, what they should optimize, and what tradeoffs they should make in daily work. This sentence shows the leader’s first responsibility in OKR operations. It is not to make people write more goals, but to decide what to choose and what to set aside during this period.

Atlassian also recommends 1–3 Objectives and 3–5 Key Results per Objective when setting OKRs. The meaning of these numbers is not simply a writing rule. Without reducing the number of goals, priorities do not emerge. In an organization that says every goal is important, there is no goal that is actually important.

The question leaders should ask is not “Should we include this goal too?” but “If we include this goal, what must we remove?” If no goals are deleted in an OKR meeting, the strategic conversation is not yet over. Employees understand priorities not by looking at the list of goals the leader approved, but by looking at the list of work the leader decided to give up.

An OKR check-in is not a reporting meeting but a conflict-resolution meeting

Atlassian recommends that teams score, analyze, and summarize OKRs every month. It also explains that regular and visible progress checks strengthen accountability and momentum. In many organizations, however, check-ins turn into reporting meetings. Each team states its progress rate, and the leader concludes by saying everyone should work harder. With this approach, OKRs struggle to become an operating rhythm.

The Google playbook says that if a team determines it cannot achieve a committed OKR, it should escalate immediately. The more important point is the purpose of escalation. It is described as a process that enables executives to create options and resolve conflicts when disagreements over priorities, shortages of time, people, or resources, or disagreements over the goal itself arise.

Therefore, the question a leader should ask in an OKR check-in is not the progress number. “What obstacle is preventing the goal from being achieved?” “Are dependencies with other teams still unresolved?” “Are resources insufficient, or has the goal’s priority become lower?” “If we do not adjust now, what loss will grow next month?” Without these questions, a check-in ends as a reading of reports.

For example, when a product-launch OKR is delayed, if the leader only says, “Raise progress to 80% by next month,” the check-in becomes a motivation meeting. By contrast, if the group separates whether the bottleneck is legal review, development capacity, or excessive expansion of sales requirements, and adjusts priorities on the spot, the check-in becomes an operating meeting. OKR leadership is not a technique for pressuring employees; it is a technique for turning conflict into decisions.

In cross-team OKRs, leadership is not a slogan about collaboration but the design of accountability

The Google playbook explains that cross-team OKRs are appropriate when important projects require contributions from multiple groups. In such cases, every group that must materially participate should be included in the OKR, and each group’s contribution should appear explicitly in that group’s OKRs.

This principle is also important in cross-department collaboration in Korean companies. Many organizations write “strengthen collaboration” as a goal, but in reality it is not clear which department must provide which output by when. If a goal requires marketing, sales, product, HR, and data teams to move together, each organization’s OKRs must show the responsibilities connected to the others. Otherwise, a shared goal becomes a goal that no one takes responsibility for until the end.

A leader is not someone who merely declares cross-team OKRs, but someone who designs the connection structure. It is not enough to create a meeting body. Leaders must disclose the dependency list, decide who to escalate to when bottlenecks arise, and align the success criteria for the shared goal in a common language. Collaboration does not work through good attitudes alone. Collaboration needs lines of accountability and authority to coordinate.

If committed and aspirational are not distinguished, leaders send the wrong signal

The Google playbook distinguishes between committed OKRs and aspirational OKRs. A committed OKR is a goal promised for achievement, and the expected score is 1.0. If it is not achieved, an explanation and retrospective are needed. By contrast, an aspirational OKR is a goal that stimulates higher challenge and innovation. Even with the same achievement rate, the interpretation should be different.

If leaders do not make this distinction, employees receive confusing signals. If a leader asks for stretch goals but criticizes low achievement rates, only safe goals will come up in the next quarter. Conversely, if underachievement of promised goals is treated only as “fine because we challenged ourselves,” execution accountability weakens. A leader who handles both types of goals with the same expression blurs the language of OKRs.

Leaders must make the type clear from the moment they approve a goal. Is this goal a commitment that must be achieved, or a challenge that pushes the organization further? If it is a committed goal, resources and priority must be guaranteed. If it is an aspirational goal, the possibility of failure should be allowed, but what will be learned must be defined. Only when this distinction exists do employees set goals honestly.

HR should provide leaders with operating questions, not just OKR forms

One reason OKR training fails is that it remains training on forms. It is not enough to explain how to write Objectives and how many Key Results are appropriate. Training must connect to what questions leaders should actually ask in meetings and what decisions they should make.

HR can provide leaders with at least four operating questions. First, what will we give up this quarter? Second, if this OKR fails, where is the most likely bottleneck? Third, are the parts that require other teams’ contributions explicitly stated in each team’s OKRs? Fourth, is this goal committed or aspirational? When these questions are repeated in meetings, OKRs become not a document but an operating language.

OKR leadership is not a matter of charisma or encouragement. It is a matter of management capability: narrowing priorities, making uncomfortable tradeoffs visible, escalating conflicts at the right time, and designing accountability between teams. Employees do not move simply by hearing goals. They move by watching what leaders choose and what they adjust. For OKRs to become an operating system for performance management, the leader’s role must also change from goal manager to designer of the execution structure.

The question companies encounter most quickly after introducing OKRs is evaluation and compensation. “Should OKR achievement rates be reflected in HR evaluations?” “If they are not connected to compensation, won’t employees fail to take them seriously?” “Conversely, if they are connected to compensation, won’t stretch goals disappear?” These questions are hard to avoid in OKR operations.

However, it is risky to rush to a single answer. If OKRs are completely separated from evaluation, execution accountability may weaken. Conversely, if OKR achievement rates are inserted into the compensation formula, employees will choose safe goals. This article is not legal or labor advice; it addresses the boundary line for linking evaluation and compensation from the perspective of HR operating design. The issue is not whether to reflect OKRs, but which OKRs should be interpreted through what evidence.

The moment OKRs enter the compensation formula, the nature of the goal changes

What Matters compares OKRs with MBO and explains that OKRs operate on a quarterly cadence and spread with a philosophy separated from compensation, unlike annual management by objectives. What matters here is not the simple rule that “OKRs must never be connected to compensation.” It is that the purpose of OKRs is not to produce evaluation scores, but to align priorities and execute strategy.

When a goal enters the compensation formula, employee behavior changes. The incentive to choose safe goals grows stronger than the incentive to choose stretch goals with a lower probability of achievement. Collaborative goals can turn into disputes over the allocation of individual responsibility. Even if strategy changes during the quarter, goals become difficult to revise, and employees try to lower initial goals to avoid a record of underachievement.

This change is a natural response. In a system where compensation is at stake, people behave according to the system. Therefore, whether to put OKRs into the compensation formula is not simply a matter of HR preference. It is a design choice connected to what the organization wants to encourage among challenge, learning, collaboration, and strategic adjustment.

Distortion occurs when committed OKRs and aspirational OKRs are handled on the same scorecard

Google’s OKR playbook distinguishes committed OKRs from aspirational OKRs. It explains that committed OKRs are close to goals whose achievement has been promised, and that the expected score is 1.0. A score below 1.0 requires explanation. By contrast, aspirational OKRs are closer to goals that stimulate innovation and challenge. They have the nature of attempting greater change before the path is fully confirmed.

If this distinction is ignored in evaluation, distortion occurs. Underachievement of a committed OKR can be a signal to review execution accountability, resource allocation, and failure to adjust priorities. But if the low achievement rate of an aspirational OKR is penalized in the same way, stretch goals disappear. Employees choose only goals favorable to evaluation, and OKRs become a list of safe promises rather than a language for innovation.

The same applies to cross-team OKRs. If goals for which several departments share responsibility are simply allocated as individual achievement rates, defensive responsibility grows stronger than collaboration. Shared goals should be reviewed together with who took which part, what dependencies existed, and what leaders adjusted. The moment all OKRs are handled with one identical scorecard, the strengths of OKRs disappear and only evaluation administration remains.

What can be used in evaluation is interpretable evidence, not achievement rates

It is also difficult to completely ignore OKR results in evaluation. If an employee held and executed an important goal throughout the quarter, the process and the result become important material for the performance conversation. However, what can be used in evaluation is interpretable evidence rather than a simple achievement rate.

The Google playbook says that Key Results should describe outcomes, not activities. This principle is also important in evaluation conversations. “Conducted training three times” is activity evidence. “Within 60 days of assigning new leaders, the rate of 1:1 feedback to team members increased from 40% to 85%” is outcome evidence. What can be referenced in evaluation is closer to the latter.

The explanation that problems with committed OKRs should be escalated quickly is also worth noting. More important than underachievement itself is what judgment and adjustment occurred when signals of underachievement appeared. When a schedule was delayed, it is necessary to check whether the leader adjusted resources, rearranged priorities, and resolved goal conflicts. A performance conversation must read the accountability structure behind the number rather than the achievement-rate number itself.

Korean companies need to design a third way between separation and reflection

In Korean companies, sensitivity around evaluation and compensation is high. The stronger an organization’s experience of connecting goal achievement rates with evaluation grades, the more quickly OKRs are also read as evaluation forms. In such organizations, even if HR declares that “OKRs are unrelated to evaluation,” employees do not easily believe it. Conversely, if HR says that “OKR achievement rates will be reflected in evaluation,” stretch goals quickly decrease.

Therefore, the realistic choice is a third way between complete separation and direct reflection. It is a method of using OKR achievement rates as reference material for performance conversations rather than directly inserting them into compensation formulas. Even then, interpretation rules by goal type are needed. Committed OKRs look at fulfillment of promises and execution accountability. Aspirational OKRs look at attempts, learning, and market or organizational signals. Cross-team OKRs look at collaboration structure and coordination accountability.

HR should leave these principles in writing. It must decide which OKRs are subject to evaluation reference, which OKRs are viewed only as learning records, what evidence will matter more than achievement rates, and whether there is or is not any disadvantage when goals are revised midway. If OKRs are operated in an ambiguous state, employees behave in the most conservative way.

Operating examples can be divided into three. First, committed OKRs are viewed as promised execution results, and reasons for underachievement and coordination accountability are checked in the performance conversation. Second, aspirational OKRs are evaluated by the hypotheses attempted, the market, customer, and organizational signals learned, and the selection criteria for the next quarter rather than by achievement rates. Third, cross-team OKRs are reviewed by looking together at each department’s promised contribution and leaders’ coordination behavior rather than splitting them into individual scores. OKRs must be distinguished this way so they are not flattened into a compensation formula.

What HR must define is not the compensation formula, but the boundary of the performance conversation

When designing the relationship between OKRs and evaluation and compensation, the first thing HR must define is not the formula. More important is the boundary of the performance conversation. HR must distinguish what will be asked in OKR reviews, what will be interpreted in evaluation interviews, and which materials alone will be used in compensation decisions.

First, OKR reviews should address progress and the need for adjustment. Second, performance interviews should look not only at OKR results but also at role expectations, collaboration, capability, and contribution to the organization. Third, compensation decisions should be explained separately within the organization’s compensation philosophy and criteria for role, grade, market value, and performance contribution. Without these boundaries, OKRs become an overloaded system that must explain everything.

Whether OKRs should be reflected in evaluation is not a matter that can end with a simple yes or no. Even if they are reflected, achievement rates should not be converted directly into scores; even if they are separated, execution accountability should not disappear. The core is to design OKRs so that they become evidence for performance conversations without killing challenge and learning. If this balance cannot be created, OKRs shrink into a supplementary item in the evaluation system. If the balance can be created, OKRs can become a performance-management language that works throughout the quarter, not only during evaluation season.

It is not unfamiliar to hear that an organization did not change even after introducing OKRs. A company-wide briefing is held, departmental Objectives are entered, and a quarter-end review schedule is set. Yet as time passes, employees accept OKRs as just another evaluation document. Leaders copy existing KPIs into a different template, and HR manages input rates and submission rates.

If OKR failure is explained only as “there were too many goals,” the core issue is missed. Having too many goals is a problem, but the bigger problem is that accountability becomes blurred. If the organization has not decided what should be treated as the top priority, which results will be recognized as real change, who will coordinate conflicts between departments, and who will reallocate resources when signs of underachievement appear, OKRs become a reporting form rather than a management language.

The first failure begins the moment OKRs are introduced only as a new template

Google’s OKR playbook describes poorly written or poorly managed OKRs as a “waste of time” and an “empty management gesture.” By contrast, it explains that well-run OKRs make clear what teams should consider important, what they should optimize, and what tradeoffs they should make in daily work.

This sentence shows the starting point of OKR failure clearly. OKRs are not a template but a language of choice and coordination. If a new template is created while the existing meeting style, leaders’ decision-making style, and the way evaluation and rewards are interpreted remain unchanged, OKRs are immediately absorbed into the existing system. From employees’ point of view, it is simply one more goal management sheet with a different name.

In Korean companies, this failure often appears under the name of “company-wide rollout.” Without a sufficient pilot, all departments are required to enter OKRs, and system registration rates are treated as implementation results. But input rate is not the outcome of OKRs. More important signals are whether OKRs actually reduced priorities, revealed conflicts between departments, and led leaders to change resource allocation decisions.

The second failure comes from filling KRs with lists of activities

The Google playbook emphasizes that Key Results should describe outcomes, not activities. It warns that KRs containing words such as consult, help, analyze, and participate may be signals that activities are being described. In HR practice, expressions such as “conduct training,” “hold interviews,” “review the system,” and “run workshops” are close to this pattern.

Activities are necessary. But activities alone do not show whether change has occurred. For example, “conduct three manager training sessions” shows what the training team did. But it does not show whether managers’ feedback behavior changed, whether team members’ understanding of goals increased, or whether the quality of performance conversations improved. When KRs are filled with activity lists, OKR reviews become meetings that check whether tasks were executed.

Good KRs ask about the change after the activity. “Increase the rate of one-on-one feedback with team members within 60 days of new leader placement from 40% to 85%” is closer to an OKR than “conduct three training sessions for new leaders.” “Increase the first response rate from candidates for critical roles from 18% to 28%” is more outcome-centered than “publish employer branding content.” Without this distinction, OKRs become a system that favors teams that write down a larger volume of work.

The third failure occurs when low-value objectives are packaged as high achievement rates

The Google playbook presents the trap of Low Value Objectives. If achieving an objective does not clearly create user value or economic value, then even a high score does not mean much for the organization. This warning applies directly to HR goals as well.

Suppose the HR department sets “complete the revision of the evaluation form” as an Objective. The form can change. But if HR cannot confirm whether the quality of evaluation conversations improved, whether goal adjustment became faster, or whether consistency in managing low performers increased, it is difficult to say that the work connected to organizational value. “Complete system revision” can become a goal with a high completion rate but low value.

In performance management, a high achievement rate is not always a good signal. The rate may be high because the goal was easy, or the score may be high because outputs unrelated to real value were produced. During OKR reviews, HR should ask not only “was it achieved” but also “who experiences what value if it is achieved.” Without this question, OKRs create documents that look like performance rather than creating performance.

The fourth failure lies in never deciding who owns a shared objective to the end

The Google playbook explains that cross-team OKRs are appropriate when an important project requires contributions from multiple groups. At the same time, it says that all groups that must substantially participate in the OKR should be included, and that each group’s contribution should be specified in its own OKR. A shared objective is not “let’s do well together”; it must reveal each party’s accountability.

This is also where OKRs become unstable in Korean companies. Goals such as improving customer experience, advancing onboarding, retaining key talent, and leadership transition cannot be achieved by HR alone. Business leaders, executives, finance, IT, and communications teams must move together. But if each organization’s contribution and decision-making authority are not written into the shared objective, the goal becomes everyone’s work and no one’s work at the same time.

A shared OKR needs three things. First, it must specify which organizations must participate. Second, it must show how each organization’s KRs connect to the overall Objective. Third, it must define who has final coordination authority when goal conflicts arise. Without these three elements, a cross-team OKR becomes a device for avoiding accountability rather than a tool for collaboration.

The fifth failure hardens the moment check-ins turn into reporting meetings

Atlassian’s OKR guide suggests 1 to 3 Objectives and 3 to 5 KRs for each Objective, and proposes a flow for regularly checking, analyzing, and summarizing progress. What matters more than the numbers themselves is the rhythm. OKRs do not exist to assign scores at the end of the quarter; they exist to adjust priorities and resource allocation during the quarter.

The Google playbook also explains that when problems arise in committed OKRs, they should be escalated quickly. The point is that raising issues with schedule, priority, or resource allocation is not merely acceptable but necessary. From this perspective, an OKR check-in is not a reporting meeting but a coordination meeting.

In Korean companies, check-ins often turn into reporting meetings. Owners explain progress rates, and leaders point out lagging items. But resource allocation does not change, and priority conflicts remain as they were. In this state, employees have no reason to update OKRs honestly. For check-ins to work, “what should we adjust” must come before “why is it late.”

The issue in the next installment is how far evaluation and rewards should be connected

Many scenes of OKR failure ultimately return to the problem of evaluation and rewards. If employees feel that goals are directly converted into evaluation scores, they choose safe goals. Shared objectives turn into disputes over individual accountability, and aspirational goals disappear. Conversely, if OKRs are completely separated from evaluation, they can look like a campaign with weak execution accountability.

So the next question is not “should OKRs be reflected in evaluation.” The more precise question is “which OKRs should be interpreted in what way.” Committed OKRs are closer to promised execution accountability. Aspirational OKRs have a strong character of learning and challenge. Cross-team OKRs require looking at collaboration and coordination accountability together. If the three types are handled with the same scorecard, OKRs are likely to fail.

To prevent OKR implementation failure, HR must design operating accountability before designing templates. Reducing the number of goals alone is not enough. HR must make people write outcomes rather than activities, filter out low-value goals, specify accountability for shared objectives, and turn check-ins into coordination meetings. Only then do OKRs become a performance management language for confirming what the organization is actually changing, rather than a reporting document.