KPIs and Business Metrics

Chapter 1.4 · KPIs and Business Metrics

Business leaders often ask for “the KPIs” as though every number has a single obvious meaning. In practice, the same metric name may refer to several definitions. Revenue might be measured before or after returns; conversion might use visits or unique people; retention might describe customers or subscriptions. Good analytics begins by specifying exactly what is counted and why.

A metric is a quantitative measure. A key performance indicator (KPI) is a selected metric used to monitor progress toward a particular objective. Not every count on a dashboard is a KPI. A well-chosen KPI has a defined owner, period, population, numerator, denominator, source and interpretation.

1.4.1 Revenue, cost and profit measures

Revenue measures the value of sales according to an agreed accounting definition. In this learning dataset, net revenue means the recorded sales value after discounts for completed orders, excluding taxes. For real reporting, the business must decide how returns, cancellations and tax treatment are handled and reconcile its rules with finance.

Gross profit = Net revenue − Cost of goods sold. The gross profit margin = Gross profit ÷ Net revenue × 100%. This describes the share of revenue remaining after the defined product costs, before other operating expenses.

Contribution margin = Net revenue − Variable costs. Variable costs may include cost of goods sold, transaction fees and variable fulfilment expenses depending on the approved definition. Gross profit and contribution margin are not interchangeable; the latter depends on which costs vary with activity.

1.4.2 Growth rates and percentage-point changes

Period-over-period growth = (Current period − Prior period) ÷ Prior period × 100%. Meridian's revenue changed from ₹10,00,000 in August to ₹8,00,000 in September: (₹8,00,000 − ₹10,00,000) ÷ ₹10,00,000 × 100 = −20%.

Growth is undefined when the prior period's denominator is zero; it is not automatically “100%”. When comparing two percentages, the absolute difference is measured in percentage points. A conversion rate changing from 5.00% to 3.67% has fallen by 1.33 percentage points. Its relative percentage decrease is approximately 26.6%.

1.4.3 Conversion and customer journey metrics

Conversion rate = Number of desired outcomes ÷ Eligible opportunities × 100%. The numerator and denominator must refer to compatible populations and periods. An online order divided by a store footfall count is not a meaningful e-commerce conversion rate. Teams must also decide whether they are measuring order conversion per session or purchaser conversion per user.

Meridian records 300 online orders and 6,000 website sessions in August, giving 5.00%. September records 220 online orders and 6,000 sessions, giving approximately 3.67%. This example uses sessions as the traffic denominator; a unique-visitor metric would require a different denominator. The lower rate is a useful signal for investigation, not proof that the checkout system caused the decline.

1.4.4 Retention, churn and customer satisfaction

Customer retention measures how many members of a defined starting group remain active after a stated period. One useful cohort definition is: Retention rate = Starting customers who remain active at period end ÷ Starting customers × 100%. New customers acquired during the period are not included in that denominator. This cohort definition must be agreed because different organizations use different activity rules.

Churn rate measures the share of a starting population lost during the period. Under compatible and exhaustive definitions, churn is 100% minus retention. For example, if 200 customers were active at the start, 170 remain active and 30 do not, the retention rate is 85% and churn is 15%. Customer reactivation and overlapping cohorts may require more careful treatment.

Customer satisfaction (CSAT) commonly means the share of valid survey responses meeting a selected satisfaction threshold. If 84 of 100 valid responses count as satisfied under the agreed rule, CSAT is 84%. Always disclose question wording, scale, response count and possible response bias; a small or self-selected sample may not represent the customer base.

1.4.5 KPI trees and analytical decomposition

A KPI tree separates a high-level measure into drivers. A simple retail identity is Revenue = Order count × Average order value, assuming both metrics use consistent definitions. Order count can be examined by channel; online orders may also be expressed as traffic × conversion rate, under the correct traffic definition.

Meridian's order count decreased from 500 to 400 while average order value stayed at ₹2,000. That decomposition suggests checking order volume first. Further segmentation reveals that online orders fell from 300 to 220 and store orders from 200 to 180. Each number is informative because its denominator, period and grouping are explicit.

1.4.6 Building a KPI dictionary

Each KPI definition should also state units, filters, reporting cut-off, authoritative system and how exceptions are resolved. If two dashboards disagree, the dictionary provides a shared basis for reconciliation.

1.4.7 Chapter summary

KPI quality depends on definition quality. Define the population and denominator before calculating a rate; distinguish gross profit from contribution; use consistent dates; and explain limitations. A useful KPI tree can organize an investigation, while an owned KPI dictionary helps multiple teams make comparable decisions.