← All writing
Tracking results · May 17, 2026 · 5 min read

What Is a KPI and How Do You Determine Them?

By John Irving

KPI stands for Key Performance Indicator. Most people know the acronym. Fewer people pay attention to what the word "key" is actually doing in that phrase.

Walk into almost any team meeting and you'll hear KPIs mentioned. They show up in slide decks, quarterly reviews, and job descriptions. Managers talk about "moving the needle" on their KPIs. Executives demand dashboards full of them. And yet, for something so universally referenced, surprisingly few people can answer a simple question: what is a KPI, really?

Not the textbook definition. The actual thing. What makes a metric a key performance indicator versus just a number someone is tracking because it's easy to track?

Getting this distinction right changes how you run your team. Getting it wrong means you spend months measuring things that feel productive but tell you nothing useful.

The Definition People Usually Get Wrong

KPI stands for Key Performance Indicator. Most people know the acronym. Fewer people pay attention to what the word "key" is actually doing in that phrase.

A KPI is not just any metric. It is a specific measure tied directly to a strategic outcome that matters to your business. The "key" part means it should be one of a small number of measures that, if they move in the right direction, signal that something meaningful is working. If a number going up or down doesn't change what you do or tell you if you're driving towards something that matters, it is not a KPI. It is just a data point.

Here is a cleaner way to think about it: a KPI answers the question, "How do we know if we are winning at the thing we said we were trying to do?" If a metric does not answer that question directly, it belongs in a different category.

There is also a distinction worth making between KPIs and metrics more broadly. Metrics are measurements. KPIs are the subset of metrics that are directly connected to strategic goals and used to guide decisions. Every KPI is a metric, but not every metric deserves to be a KPI.

Where KPIs Came From

Understanding the history of KPIs helps explain why so many organizations misuse them today. The concept did not spring out of nowhere. It evolved over about 70 years of management thinking, and each era left its fingerprints on how we use these tools.

The early roots: Frederick Taylor and scientific management

The idea that you could measure work to improve it goes back to Frederick Winslow Taylor's Principles of Scientific Management, published in 1911. Taylor believed that if you broke work down into measurable tasks and tracked output carefully, you could optimize almost any process. His approach was built for factory floors, but the underlying logic, that measurement drives performance, became the foundation for everything that came after.

Peter Drucker and Management by Objectives

The real turning point came in 1954, when Peter Drucker published The Practice of Management and introduced Management by Objectives (MBO). Drucker's insight was that organizations should start with goals and then figure out what to measure, rather than the other way around. MBO asked managers and employees to jointly agree on specific objectives and then track progress toward them. The KPI, as we use it today, is a direct descendant of this idea.

Drucker also warned against something that still plagues teams today: measuring what is easy rather than what matters. His concern was that once you start tracking a number, people optimize for that number, whether or not it reflects real progress. This is the Goodhart's Law problem, which states that once a measure becomes a target, it ceases to be a good measure.

The Balanced Scorecard

In 1992, Robert Kaplan and David Norton introduced the Balanced Scorecard, which expanded KPIs beyond financial metrics. Before the Balanced Scorecard, most organizations measured success almost entirely through revenue, profit, and cost. Kaplan and Norton argued that financial numbers alone were lagging indicators: by the time they moved, the underlying problems or opportunities were already old news.

Their framework organized KPIs into four perspectives: financial, customer, internal processes, and learning and growth. The idea was that a healthy organization should be able to show progress across all four areas, not just the income statement. This is still one of the most influential frameworks in use today.

OKRs and the modern era

The most recent evolution is OKRs (Objectives and Key Results), popularized by Intel and later by Google. OKRs are a close cousin of KPIs, but they are built for faster-moving environments where teams set ambitious quarterly goals rather than tracking steady-state performance. Where a KPI asks "are we at the level we need to be," an OKR asks "are we making progress toward a stretch goal?" They serve different purposes, and many organizations use both.

In practice, KPIs and OKRs operate at different layers. KPIs are ongoing: they measure the health of something you always need to maintain, like retention, margin, or response time. OKRs are time-boxed: they focus a team on a specific ambition for a quarter or a year, then reset. A common pattern is for a KPI that drops below an acceptable threshold to become the motivation for an OKR. Once the OKR is achieved and the KPI recovers, the OKR goes away and the KPI continues doing its job in the background. The failure mode is treating them as duplicates and running both without a clear distinction between them.

OKRs are worth a deeper look on their own — we'll cover how to build and run them in a follow-up piece.

What Gets People Into Trouble

Given this history, why do so many teams still struggle with KPIs? A few patterns come up again and again.

Tracking vanity metrics

A vanity metric is a number that looks good but does not inform a decision. Website visitors, social media followers, emails sent, and gross sign-ups are common examples. These numbers tend to go up over time regardless of whether the business is actually getting healthier. They feel productive to track because they are easy to move, but they rarely tell you whether you are achieving something meaningful.

A useful test: ask whether a metric going up would actually change what your team does. If the answer is "not really," that metric is probably a vanity metric in disguise.

Tracking too many things

There is a common instinct to track everything, especially when data is easy to pull. The result is a dashboard with 20 or 30 metrics, which means no one knows which ones actually matter. A useful rule of thumb is that any team should have no more than five to seven true KPIs at any given time. If everything is a priority, nothing is.

Missing the three elements that make a KPI usable

A number alone is not a KPI. For a metric to be genuinely useful, it needs three things: a target (what "good" looks like), a trend (how it is moving over time), and an owner (a specific person responsible for moving it). Without all three, a KPI is just a data point someone put on a slide.

Confusing activity with outcome

Teams often default to measuring inputs: calls made, reports written, features shipped. These are easier to track than outcomes, but they do not answer the real question, which is whether the work is producing results. A sales team might hit their call volume KPI every week and still miss their revenue goal. The activity metric gave them a false sense of control.

The Process: How to Actually Think About KPIs

Setting good KPIs is less about picking the right metrics from a list and more about thinking through a specific sequence of questions. Here is a practical process for managers and team leads.

Step 1: Start with the outcome, not the metric

Before you name a single KPI, write down what success looks like for your team in clear language. Not "improve customer satisfaction," but "customers renew at a rate above 85% within 30 days of their contract end date." The more specific the outcome, the easier it is to identify the right measure.

Step 2: Ask what would have to be true for that outcome to happen

Work backwards from the outcome. If your goal is high renewal rates, what drives renewals? Product adoption, customer support response time, account manager check-ins? This step maps the levers that actually move the outcome, and those levers often become your KPIs or the leading indicators underneath them.

Step 3: Distinguish leading from lagging indicators

Lagging indicators tell you what happened: revenue, churn rate, net promoter score. Leading indicators predict what will happen: product login frequency, support ticket resolution time, onboarding completion rate. Good KPI sets include both. Lagging indicators tell you whether you won. Leading indicators tell you whether you are on track to win.

Step 4: Apply the decision test

For each candidate KPI, ask: if this number moved significantly in either direction, would it change what we do? If the answer is yes, it belongs on your list. If the answer is "we would look into it" or "we would mention it in a meeting," it does not.

Step 5: Assign an owner and a target

Every KPI needs a single owner: not a team, not a department, one person. That person is responsible for understanding why the number is moving, communicating what they plan to do about it, and making decisions based on what they see. Pair the owner with a clear target so there is no ambiguity about what success looks like.

Step 6: Review the set regularly

KPIs should not be permanent fixtures. They reflect the current priorities of your business, and priorities change. A good cadence is to formally review your KPI set every quarter. Ask whether the goals behind each KPI are still the right goals, whether the metrics are still the best proxies for those goals, and whether the targets still reflect what is achievable and meaningful.

A Useful Frame: The KPI as a Question

One of the most practical ways to think about KPIs is to treat each one as a standing question your team is trying to answer. Instead of "our KPI is customer churn rate," think of it as "we are constantly asking: are we retaining the customers we worked hard to acquire, and are we doing it at a rate that makes the business sustainable?"

Framed this way, a KPI is not just a number on a dashboard. It is a lens that keeps your team focused on something that matters, and a signal that tells you when to act. The metric is the answer. The KPI is the question.

Most teams get into trouble because they accumulate answers without ever deciding which questions they are really trying to ask. Fixing that, starting with fewer, sharper, more deliberate KPIs, is usually the most impactful management change a team can make.

Where to Start

If you are sitting with a list of KPIs your team currently tracks, try this: go through each one and ask whether it passes the decision test. Can you point to a recent example where that number moving led to a specific action? If you cannot, that metric may not deserve the "key" in its name.

Then start over with the outcome. What does your team actually exist to do? What would have to be true for you to say, at the end of the year, that this team succeeded? Write that down first. Everything else flows from there.

One area where the traditional KPI framework gets stress-tested is early-stage product development. AI has compressed the build and validation cycle dramatically — teams can go from idea to working software to user feedback in days rather than months. At that speed, quarterly KPIs are too slow to be useful. What you're really tracking is whether your assumptions are holding up in near real-time: are users engaging the way you expected, and is the problem you set out to solve actually the problem they have? The measurement discipline is the same, start with the outcome, track what tells you if you're moving toward it, but the cadence and the nature of the signals look very different.

Sources

History and frameworks

KPI definition and selection

Vanity metrics