Digital MarketingModule 4: Measurement and economicsLesson 10 of 17
Course progress53%
19 min lesson · Updated August 2026
What are marketing KPIs and analytics?
Marketing analytics collects and interprets evidence about audience and campaign behavior; a KPI is a deliberately chosen measure that helps a specific decision about an important objective.
What you will learn
By the end, you will understand:
Build a measurement hierarchy from objective to diagnostic signals
Define metrics with owner, formula and data limits
Test implementation, privacy and real-world reconciliation
Visual explainer
See the idea clearly.
Business objective sits above outcome KPI, journey indicators, channel diagnostics and quality/risk measures; data sources connect through consent and validation to decisions.
Not every metric is a KPI
Analytics tools expose hundreds of numbers. A KPI is important because it helps judge progress or trigger action for a defined objective. Other metrics diagnose why it changed.
A dashboard with every available metric creates attention without decision. Start with the decision and evidence required.
A measurement hierarchy
Level
Example
Objective
Increase profitable first purchases from eligible new customers.
Outcome KPI
New-customer contribution after acquisition cost.
Journey indicator
Qualified checkout completion.
Channel diagnostic
Relevant click-through and landing-page conversion.
Quality/risk
Returns, complaints, tracking errors and consent failure.
Define every important metric
Purpose/decision
Formula
Numerator/denominator
Eligible population
Time zone/window
Source/system
Event trigger
Deduplication
Attribution model
Observed versus modeled
Owner
Known limitations
Implementation must reflect real success
A form conversion should fire after server-confirmed success, not merely on button click. Purchases need transaction IDs and correct values/currency. Internal/test traffic and duplicate events can inflate performance.
Test in the browser, analytics debugger/realtime and downstream reports, then reconcile with CRM/order records. A green tag does not prove the business outcome.
Quantitative and qualitative evidence work together
Quantitative
Qualitative
Operational
Experimental
Shows scale, rate, distribution and change across defined events.
Interviews, usability, calls and feedback help explain motives, confusion and context.
CRM, fulfilment, returns and support show whether marketing promises became real value.
Controlled tests can strengthen causal inference when design and sample are appropriate.
Protect people and consent choices
Do not send names, emails, phone numbers or free-text personal information to Google Analytics. Collect only necessary events and restrict access/retention.
Consent requirements vary. Consent Mode communicates choices to Google tags and adjusts behavior; the site remains responsible for obtaining and honoring choices and controlling all tags.
Use uncertainty in reporting
Sampling, modeling, consent loss, attribution, identity gaps, offline activity, bot traffic and data delay affect results. Label estimates and avoid presenting platform precision as certainty.
A KPI can move because of price, product, operations or market change—not marketing alone. Add annotations and alternative explanations.
Real-world example
Example: fixing a false lead KPI
Example
A company counted form-button clicks as leads. QA shows validation errors and abandoned requests. It changes the event to server-confirmed submission, adds submission-ID deduplication and reconciles weekly with CRM records. Reported volume falls but becomes useful.
Try this
Write one metric contract
For your primary KPI, document formula, event, source, eligibility, time, deduplication, attribution, owner and three limitations. Test one successful and one failed user path.
Common questions
Questions beginners ask.
What is a KPI?
A key performance indicator selected to help judge or decide against an important objective.
What is marketing analytics?
Collection and interpretation of marketing-related evidence to understand behavior, performance and decisions.
Is every dashboard number a KPI?
No. Many metrics are supporting diagnostics.
What is a vanity metric?
A number that looks impressive but offers little decision value without context, such as raw views unrelated to audience or outcome.
When should a lead event fire?
After the valid enquiry is durably accepted, not simply when a button is clicked.
Why reconcile analytics with CRM?
To find duplicates, missing events and differences between tracked actions and real business records.
Can GA4 receive email addresses?
Google policies prohibit sending personally identifiable information such as email addresses to Analytics.
What is modeled data?
Estimated activity produced under documented modeling methods when direct observation is incomplete; it is not exact individual reconstruction.
