KPI · Key Performance Indicator
Key Performance Indicator
What is a KPI?
- The key metric for goal achievement
- Leading + lagging together
- When the metric becomes the goal, it gets distorted (Goodhart’s law)
The first thing KPIs change is the language of meetings. A report like “Training feedback has been good lately” becomes “Completion rate 62%, 18 points below target,” and from that moment the discussion shifts from arguing over impressions to finding causes. The real function of a KPI is not evaluation but setting a baseline for conversation. When departments compete over where to put more budget and headcount, letting the numbers speak first reduces cross-department politics.
Success depends less on the metric’s name than on its “metric definition sheet.” Even for the same “on-the-job application rate,” the number can differ by a factor of two depending on whether the denominator is course enrollees or all target employees, whether the data comes from LMS logs or surveys, and whether it is aggregated monthly or quarterly. So put four things in writing first: the formula, the source, the cycle, and who is accountable. The count matters too. Unless you cap it at 3–5 core metrics per team and one owner per metric, your dashboard quickly turns into a bulletin board of 20 numbers.
If you only look at lagging indicators, the moment to act has already passed. Revenue, completion rate, and turnover are outcomes. To intervene midstream, pair them with leading indicators that move one to two months ahead: number of consultations, the share of team leaders holding 1-on-1s, and week-1 learning login rate. Always set an intervention threshold for each leading indicator. Without a rule like “If the week-1 login rate is below 70%, notify the team leader that week,” even a leading indicator ends up as a number in an after-the-fact report.
The most common failure happens the moment a KPI is converted directly into a performance-review score. When hitting the target becomes the grade, people set targets low (defensive goal-setting), rush through token course completions at year-end, and slip in a “please give us a 5” just before the survey. That is how you end up with rising numbers and an unchanged workplace. Adding metrics the owner can’t control produces the same result. The organization has to agree up front on the principle that a KPI is a ruler for measuring performance, not a report card for ranking people.
Seungbin JeonAn instructor who draws on psychology expertise to support communication and leadership growth in organizations
John KimInstructor John Kim, author of B2B Marketing Design and CEO of Nanum ConsultingA standardized metric concept that grew out of the measurable performance management tradition stemming from Peter Drucker’s management by objectives (MBO, 1954). It shows, in numbers, how far an organization is achieving its core goals.
Consider a hypothetical case. Company A makes industrial components and has 740 employees and annual revenue of about KRW 210 billion (≈ US$157,115,100). Five months after launching company-wide AI training, its completion rate had stalled at 41%, and it had no numbers to show what had changed on the job. The four-person training team set aside the three months of Q2 to redesign its metrics.
- At the management meeting, the head of training reported, “I can tell you the completion rate is 41%, but we have no numbers on what has changed on the job.” The CEO then made it clear: cut the metrics down to three.
- Of the 12 metrics it had been tracking, the training team kept only three: “AI training completion rate,” “number of on-the-job application cases within 3 months,” and “monthly hours saved on applied projects.” The rest were demoted to reference metrics.
- While drafting the definition sheet, the training manager wrote, “The completion-rate denominator is all 740 target employees, not just those who signed up,” and fixed the source as LMS logs only.
- They added “week-1 login rate” as a leading indicator and set up an automatic alert to the team leader that week for any team below 70%. The first response came from the production management team leader: “We’re at 48%. I’ll bring it up at the morning briefing.”
- In the quarterly review, the team asked for a one-line cause for each missed metric. 62% answered that “training hours overlap with regular shifts,” so they changed the schedule to two repeat sessions for night-shift staff.
- Three months later, the completion rate rose from 41% to 83%, and on-the-job application cases grew from 7 to 34. Across those projects, an estimated 175 work hours per month were saved.