KPI Examples: Definition, Formulas and How to Build a KPI Dashboard

What is a KPI?

A KPI (key performance indicator) is a number that tells you whether you are reaching a specific business goal. "Monthly revenue" is a metric; "monthly revenue versus a 2 million MAD target, owned by the sales director, reviewed every Monday" is a KPI. The difference is the link to a goal, a target and someone accountable.

Every KPI is a metric, but most metrics are not KPIs. A company can track hundreds of numbers; only a handful deserve a place at the top of a dashboard.

What is the difference between a KPI, a metric and an OKR?

Concept	What it is	Example
Metric	Any measured value	Number of website visits
KPI	A metric tied to a goal, with a target and an owner	Conversion rate ≥ 2.5 % this quarter
OKR	An objective plus 3–5 measurable key results, usually quarterly	"Grow online sales": reach 1,200 orders, cut cart abandonment to 60 %
Leading indicator	Moves before the result	Number of qualified quotes sent
Lagging indicator	Confirms the result after the fact	Revenue invoiced

A useful dashboard mixes both: lagging KPIs tell you where you are, leading KPIs tell you where you are going.

What makes a good KPI?

The classic SMART checklist (specific, measurable, achievable, relevant, time-bound) still works. In practice, five questions are enough:

• Is the formula written down? "Margin" can mean gross margin, contribution margin or net margin. If two people compute it differently, it is not a KPI yet.
• Where does the data come from? ERP, CRM, accounting software, a spreadsheet maintained by hand.
• Who owns it? One person or role who explains the number and acts on it.
• What is the target and the alert threshold? Without a target, a number is just information.
• Can someone act on it? If nobody changes a decision when it moves, remove it.

KPI examples by department

The table below lists common KPIs with their formula. Adapt names and targets to your sector.

Department	KPI	Formula
Sales	Revenue vs target	Revenue / Target revenue
Sales	Average order value	Revenue / Number of orders
Sales	Quote conversion rate	Quotes won / Quotes sent
Marketing	Cost per lead	Marketing spend / Leads generated
Marketing	Website conversion rate	Orders (or forms) / Sessions
Finance	Gross margin %	(Revenue − Cost of goods sold) / Revenue
Finance	Days sales outstanding (DSO)	(Accounts receivable / Revenue of the period) × Days in the period
Customer service	First contact resolution	Tickets solved at first contact / Total tickets
Customer	Churn rate	Customers lost in the period / Customers at start of period
Operations	On-time delivery rate	Orders delivered on time / Orders delivered
Operations	Stock rotation	Cost of goods sold / Average stock value
HR	Absenteeism rate	Days absent / Days worked planned
HR	Turnover	Departures in the period / Average headcount

Sector examples in Morocco

• Retail and distribution: sales per square metre, stock-outs, average basket, sales per store and per region (Casablanca-Settat, Rabat-Salé-Kénitra, Tanger…).
• Telecom and subscription services: churn, ARPU (average revenue per user), activations per channel.
• Banking and insurance: cost-to-income ratio, non-performing loan ratio, claims processing time.
• Public administration and social-security-type data: processing time per file, declarations received on time, coverage rate per region. The CNSS decision-support project is an example of indicators read by period, region and category.

How do you define a KPI? The KPI sheet

Before building any chart, write a one-page KPI sheet for each indicator and get it validated by its owner:

Field	Example
Name	On-time delivery rate
Business question	Are we keeping our delivery promise?
Formula	Orders delivered on or before the promised date / Orders delivered
Source	Order management system, table of deliveries
Grain and filters	Per order; excludes cancelled orders
Frequency	Weekly, reviewed every Monday
Owner	Logistics manager
Target / alert	Target ≥ 95 %, alert below 90 %

This sheet prevents the most common dashboard problem: two reports showing two different numbers for the "same" KPI.

How do you calculate a KPI in SQL or DAX?

Once the definition is fixed, the calculation is usually short. In SQL (PostgreSQL), the on-time delivery rate per month:

SELECT
    date_trunc('month', delivered_at)::date AS month,
    ROUND(
        100.0 * COUNT(*) FILTER (WHERE delivered_at <= promised_at)
        / NULLIF(COUNT(*), 0),
        1
    ) AS on_time_rate_pct
FROM deliveries
WHERE status <> 'cancelled'
GROUP BY 1
ORDER BY 1;

In Power BI, the same logic as DAX measures, assuming a Deliveries fact table and a date table:

Delivered Orders =
COUNTROWS ( FILTER ( Deliveries, Deliveries[Status] <> "cancelled" ) )

On-Time Orders =
COUNTROWS (
    FILTER (
        Deliveries,
        Deliveries[Status] <> "cancelled"
            && Deliveries[DeliveredAt] <= Deliveries[PromisedAt]
    )
)

On-Time Rate % = DIVIDE ( [On-Time Orders], [Delivered Orders] )

Measures like these only behave correctly on a clean model. If filters by date or region give strange totals, the issue is usually the data model; see how to build a Power BI star schema and the essential DAX formulas.

How many KPIs should a dashboard show?

For a management dashboard, 5 to 8 KPIs on the first page is a good order of magnitude. Beyond that, nobody reads them. A simple layout:

1. Top row: the main KPIs as cards, each with its value, target and trend versus the previous period.
2. Middle: one or two charts that explain the main KPI (trend over time, breakdown by region or product).
3. Bottom or second page: detail tables for people who need to investigate.

Other rules that make dashboards usable:

• One audience per dashboard: executives, sales managers and operations need different views.
• Colours carry meaning (red only for an alert), not decoration.
• Always show the comparison: versus target, versus last month, versus same period last year.
• Show the data refresh date so users know how fresh the numbers are.

Power BI has a dedicated KPI visual (value, target and trend); its behaviour is documented on Microsoft Learn. For page-by-page layout ideas, see these Power BI dashboard examples.

Common KPI mistakes to avoid

• Vanity metrics: page views or followers that look good but do not drive a decision.
• Too many KPIs: when everything is "key", nothing is.
• No single definition: finance and sales each compute margin their own way.
• Targets that nobody owns: a red KPI that nobody is responsible for stays red.
• Manual data: a KPI rebuilt by hand each month in Excel is slow and error-prone; automate the data flow once the definition is stable.
• Ignoring data quality: duplicates, missing dates or wrong codes silently distort KPIs. Basic checks, like those described in data quality and anomaly detection, catch these problems before they reach the dashboard.

Conclusion

A KPI is a number with a purpose: a goal, a formula, an owner and a target. Start with a short list per audience, write a KPI sheet for each, compute them on a clean model, and display them with a comparison so people can act. In the D&A Studio platform, for example, Power BI dashboards were built on top of explicit KPI definitions (see the project).

If you want help defining your indicators or turning them into a Power BI dashboard, see the Power BI dashboard service.