Why Most KPIs Are Lies.
Most organizations measure activity. Very few measure understanding.
Most KPIs are lies because they measure activity instead of outcomes. A team can hit every target on the dashboard while the business gets worse, because the dashboard is rewarding movement, not progress. This guide explains why this happens, names the principle behind it (Goodhart's Law), and walks through the small set of metrics that resist gaming and predict real growth.
What Is A Vanity Metric, And Why Are Most KPIs One?
A clean definition and the test that separates a vanity metric from a real one.
[ DEFINITION ]
Vanity Metric: A number that can rise reliably without the underlying business outcome improving. Vanity metrics are easy to grow, easy to report, and almost impossible to tie to revenue, retention, or a real customer result.
The fastest test is this. Ask whether the metric can double while the business gets worse. If the answer is yes, the metric is a vanity metric. Followers can double while revenue collapses. Emails sent can double while reply rates fall. Tickets closed can double while customers churn faster. The number is moving. The outcome is not.
The reason most KPIs end up in this category is simple. Activity is easy to count, outcomes are hard to attribute, and dashboards reward what can be counted today, not what matters next quarter.
Goodhart's Law: When A Measure Becomes A Target.
The principle that explains why almost every KPI eventually breaks.
[ DEFINITION ]
Goodhart's Law: When a measure becomes a target, it ceases to be a good measure. Once people are evaluated against a number, they will optimize for the number, not for the underlying outcome the number was supposed to represent.
Goodhart was writing about monetary policy in 1975, but the rule applies anywhere a number gets attached to a paycheck or a performance review. The moment "reduce ticket time" becomes a target, ticket time falls and customer satisfaction follows it down. The moment "increase meetings booked" becomes a target, calendars fill with low-quality meetings.
This is not a failure of the people on the team. It is a failure of metric design. Single-number targets without a counterbalancing quality metric will always degrade. Andy Grove called this "paired indicators" in High Output Management. Pair output with quality, pair speed with accuracy, and the gaming stops being easy.
Activity vs Outcomes: The Difference That Kills Companies.
Hitting the activity target while missing the result is the most common failure mode in modern teams.
A salesperson can hit a 100-call-per-day quota and generate zero revenue. A marketing team can double traffic and produce zero new customers. A support team can close tickets 30% faster and watch CSAT drop 20%. In every case, the activity KPI says success and the business says otherwise.
[ DEFINITION ]
Outcome Metric: A measurement tied to a change in the real world a customer or shareholder would recognize as value: revenue won, customer retained, problem permanently solved, time returned to a person.
Outcome metrics are harder to capture because they require context, attribution, and time. A meeting booked today might not become revenue for six months. A support fix today might prevent twenty tickets next year. The lag is exactly why teams default to activity metrics, and exactly why they keep losing.
How The Report Quietly Replaces The Result.
The drift from objective to KPI to meeting to performance theater.
The pattern is consistent across companies. Someone sets an objective. The objective gets translated into a KPI. The KPI becomes a weekly report. The report becomes a meeting. Eventually people start preparing for the meeting instead of solving the problem the objective was created to solve.
By the time the dashboard is green every week, the metric has survived and the objective has died. The team is still moving. The business has stopped progressing. This is what most "KPI culture" looks like in practice.
What To Measure Instead: Learning, Not Movement.
Three categories of metric that resist gaming and predict real growth.
The organizations that improve fastest are not the ones collecting the most data. They are the ones turning experience into intelligence. Three categories matter more than activity.
[ DEFINITION ]
Learning Metric: A measurement of whether this period's decisions outperformed the previous period's. Did the win rate improve. Did fewer tickets recur. Did the cost of acquiring a customer fall as the team got more reps.
Outcome metrics prove value. Learning metrics predict whether next quarter will be better than this one. Trust metrics (how much work runs without escalation, how often the team is right when it commits) are the third category, and the one most operators ignore.
The Loop That Separates Learning Teams From Busy Ones.
The teams that compound do four things on repeat. They remember what happened in enough detail to study it later. They understand why it happened, not just that it happened. They adjust their next action based on what they understood. Then they do it again. Activity dashboards skip steps two and three entirely.
A Tale Of Two Teams: 10,000 Emails vs 1,000.
The team that gets smarter beats the team that gets busier. Every time.
Team A sends 10,000 emails this quarter. Team B sends 1,000. A traditional activity dashboard rewards Team A. But Team B captures the reason every reply happened, tags every objection, and rewrites the next send based on what came back. Team A's reply rate stays flat. Team B's reply rate doubles every quarter.
Two quarters later, Team B is sending fewer emails and winning more deals than Team A. Three quarters later, the gap is not recoverable. Team B did not work harder. It got smarter, and the system kept getting smarter, and compound interest did the rest.
What To Do This Week.
Five concrete steps to retire vanity KPIs and replace them with outcome and learning metrics.
Audit every KPI on your current dashboard with one question. Can this number double while the business gets worse. Retire any metric where the answer is yes. Pair every remaining output metric with a quality counterpart (calls plus connect rate, tickets closed plus repeat-ticket rate). Add one learning metric per team: a measurement of whether this period's decisions outperformed the last. Stop attaching compensation to single-number targets. Review the new dashboard in 90 days and prune again.
Sources.
Primary research and authoritative references behind this piece.
- Goodhart's Law (original 1975 monetary policy paper, Charles Goodhart). Background on why measures stop being good measures once they become targets
- Andy Grove, High Output Management (paired indicators). Source of the paired-metric principle referenced above
- Harvard Business Review: The Trouble With KPIs
KPI Questions People Actually Ask.
Direct answers to the questions search and AI assistants surface around vanity metrics and outcome measurement.
- What is the difference between a vanity metric and a real KPI?
- A vanity metric goes up reliably without changing the outcome (followers, page views, emails sent). A real KPI moves with the result you actually care about (revenue, retention, problems solved). If the metric can climb while the business gets worse, it is a vanity metric.
- Why do so many KPIs stop working over time?
- Goodhart's Law. The moment a measure becomes a target, people optimize for the measure instead of the underlying goal. The metric stays green, the gaming gets sophisticated, and the outcome quietly degrades. This is why metrics need to be refreshed, not just tracked.
- What should I measure instead of activity?
- Measure outcomes (did the customer renew, did revenue grow, did the problem stop recurring) and learning (did this quarter's decisions outperform last quarter's). Outcomes prove value. Learning predicts whether next quarter will be better than this one.
- How many KPIs should a team actually track?
- Three to five per team is enough. More than that and people stop paying attention to any of them. Andy Grove's High Output Management argues that paired metrics (one for output, one for quality) prevent the most common gaming of single targets.