tadata
← Back to home

KPI Frameworks: Measuring What Actually Matters

#analytics#kpi#strategy#business-intelligence#management

Not everything that can be measured should be measured. KPI frameworks provide structured approaches to selecting, organizing, and acting on the metrics that drive business outcomes.

Framework Comparison

FrameworkOriginBest ForKey Idea
OKRsIntel / GoogleAlignment across teamsAmbitious objectives with measurable key results
Balanced ScorecardKaplan & NortonEnterprise strategy executionFour perspectives: financial, customer, process, learning
AARRR (Pirate Metrics)Dave McClureProduct-led growthAcquisition, Activation, Retention, Revenue, Referral
North Star MetricSean EllisFocus and simplicityOne metric that best captures value delivered to customers
HEARTGoogleUX quality measurementHappiness, Engagement, Adoption, Retention, Task success

OKRs in Practice

OKRs work when they are used for alignment, not performance evaluation. Common pitfalls:

  • Setting OKRs that are actually task lists, not outcomes
  • Making every OKR achievable (they should be stretched to 70% achievement)
  • Failing to cascade OKRs so team objectives connect to company objectives
  • Reviewing only at quarter-end instead of weekly check-ins

Structure: Objective (qualitative, inspiring) + 3-5 Key Results (quantitative, measurable, time-bound).

The North Star Metric

A North Star Metric is the single metric that best reflects the core value your product delivers to customers.

Company TypePossible North StarWhy
MarketplaceWeekly transactions completedReflects value exchange for both sides
SaaSWeekly active teams using core featureMeasures habitual value delivery
MediaTotal engaged reading timeCaptures attention and content quality
E-commerceWeekly purchases per active customerCombines frequency and conversion
FintechMonthly assets managed / transactedReflects trust and utility

The North Star is not a vanity metric. It must correlate with long-term revenue and be influenceable by the product team.

Leading vs Lagging Indicators

AspectLeading IndicatorsLagging Indicators
TimingPredict future outcomesMeasure past results
ActionabilityDirectly influenceableOutcome of many factors
ExamplesPipeline deals, feature adoption, NPSRevenue, churn rate, market share
Use forCourse correctionPerformance evaluation
RiskMay not correlate with outcomesToo late to act on

A healthy KPI system includes both. Leading indicators give you time to react; lagging indicators confirm whether your actions worked.

KPI Trees

A KPI tree decomposes a top-level metric into its component drivers:

Revenue breaks down into:

  • Number of customers x Average revenue per customer
    • New customers + Retained customers - Churned customers
      • Leads x Conversion rate = New customers
      • Existing customers x (1 - Churn rate) = Retained customers

Each leaf of the tree becomes a team-level KPI. This ensures every team understands how their work connects to company outcomes.

Setting Targets

  • Historical baseline: What was last period's performance? Targets without baselines are guesswork
  • Benchmarks: What do comparable companies achieve? Use industry reports cautiously
  • Capacity constraints: What is physically possible given current resources?
  • Stretch vs committed: Distinguish between "we expect to hit this" and "this would be exceptional"

Avoiding Vanity Metrics

A vanity metric looks impressive but does not inform decisions. Common traps:

  • Total registered users (instead of monthly active users)
  • Page views (instead of engagement depth or conversion)
  • Downloads (instead of activation or retention)
  • Social followers (instead of engagement rate or attributed revenue)

The test: "If this metric changed, would we change what we are doing?" If not, it is vanity.

Resources