Google Business Intelligence Professional Certificate
A 4-course Google certificate teaching data modeling, visualization, and dashboarding for business intelligence career roles in under 2 months.
A metric is any quantity you measure: page views, orders, refunds, average delivery time. A KPI (key performance indicator) is one of the few metrics a team has agreed are most important for judging progress toward a specific goal, usually with a target and an owner. Every KPI is a metric, but most metrics are not KPIs. A dashboard with 40 "KPIs" really has 40 metrics and no priorities.
Choosing what to measure is one of the most valuable things an analyst does, and one of the least taught. The wrong KPI can push a whole team in the wrong direction while every chart looks healthy.
An online store tracks dozens of numbers. A few of them:
| Number | Metric? | KPI? | Why |
|---|---|---|---|
| Website sessions | Yes | Usually no | Useful context, but traffic alone doesn't mean success |
| Conversion rate | Yes | Often yes | Directly tied to revenue and something the team can influence |
| Monthly revenue | Yes | Yes (company level) | The outcome the business exists to produce |
| Repeat purchase rate (90 days) | Yes | Yes (retention team) | Tied to long-term value |
| Social media followers | Yes | Rarely | Easy to grow, weakly linked to revenue |
| Average delivery time | Yes | Yes (operations team) | Drives satisfaction and repeat purchases |
Notice that whether something is a KPI depends on the goal and the team. Delivery time is a KPI for operations and background context for marketing.
Good KPI sets pair them. "Monthly revenue" (lagging) plus "checkout conversion rate" and "repeat purchase rate" (leading) tells a team both where it stands and where it's heading.
Some companies choose one north star metric: a single measure that best captures the value customers get, which the whole company aligns around. Examples are weekly active users for a collaboration app, or orders delivered on time for a delivery business. It's useful for focus, but it still needs supporting KPIs and guardrails underneath.
Vanity metrics look impressive but don't help decisions. Examples are total registered users (it never goes down), page views, raw app downloads and follower counts. The test: if this number changed, would we do anything differently? If not, it isn't a KPI.
Better versions tend to be rates and ratios: active users rather than registered users, conversion rate rather than total visits, revenue per customer rather than customer count.
Goodhart's law, in its popular form, says that once a measure becomes a target, it stops being a good measure, because people optimize the number instead of the goal behind it.
The fix is counter-metrics (guardrails): pair each KPI with a metric that would reveal gaming. Tickets closed plus reopen rate. Conversion rate plus average margin. This is the same idea as the guardrail metrics in a well-designed A/B test.
A good KPI is:
Our Power BI dashboard tutorial shows this layout in practice.
KPI design sits at the heart of business intelligence work. The Google Business Intelligence Certificate covers stakeholder requirements, metric definition and dashboard design. Browse all Business Intelligence & Reporting courses.
Is revenue a KPI or a metric? Both. It's a metric, and at company level almost always a KPI. For an individual team it may be background context rather than their own KPI.
What's the difference between a KPI and an OKR? An OKR (Objective and Key Results) is a goal-setting framework: an objective plus a few measurable key results for a period. Key results often use KPIs, but OKRs describe what you want to achieve this quarter, while KPIs monitor ongoing performance.
How many KPIs should a dashboard have? Usually three to seven for a team-level dashboard, with supporting metrics available on drill-down.
What is a vanity metric? A number that looks good but doesn't inform decisions, such as total sign-ups or page views, often because it can only go up or is only loosely linked to business results.
Metrics measure things. KPIs are the few metrics tied to a goal, with an owner and a target. Pair lagging outcomes with leading indicators, replace vanity counts with rates, guard every KPI with a counter-metric, and keep the list short. Choosing well is often worth more than any analysis done on the numbers afterwards.