Representative example 03 · Growth
Building a measurement foundation and a controlled test plan for digital advertising
This example shows how a business advertising across several channels can assess its measurement confidence and testing discipline before increasing its budget.
Starting point
Why is a decision system needed?
Platform reports can claim the same conversion in different ways, while missing tags and inconsistent campaign names make comparison harder. Budget decisions built on unreliable measurement create false certainty.
Decision question
The question the work will answer
Which part of the existing data is reliable enough to base decisions on, and which hypothesis should the next test examine with the lowest controllable risk?
Method
Four controlled stages
- 01Measurement confidence table
Measurement inventory
Platform, analytics and server-side events are compared by name, trigger condition, value, currency and deduplication.
- 02Metric dictionary
Decision metric
The primary business outcome, early signals and guardrail metrics are separated; the reporting window and data latency are documented.
- 03Approved test card
Hypothesis and test design
A single main variable, the target audience, budget limit, duration, stopping condition and how success will be interpreted are defined in advance.
- 04Decision and learning log
Reading the results
At the end of the test, data quality, external influences and uncertainty are set out; the decision to scale, repeat or stop is recorded with its rationale.
Expected deliverables
- Measurement inventory by channel and event
- Metric dictionary explaining sources, formulas and use in decisions
- Naming, UTM and campaign classification standard
- Test cards with budget limits and stopping conditions
- Decision report template that separates data quality from uncertainty
How is it measured?
- Consistency between expected and observed conversion events
- Proportion of sessions with a correctly classified source and campaign
- Proportion of tests completed in line with predefined conditions
- Business metrics such as acquisition cost and conversion value, assessed alongside attribution differences and data latency
Assumptions and trust boundaries
- Advertising platforms, analytics, conversion points and the necessary permissions can be reviewed together.
- The primary conversion and the intermediate indicators that matter to the business are defined with an authorised person.
- No ROAS, revenue or customer acquisition cost outcome is guaranteed; budget changes are subject to written approval.
- This content does not contain real advertising account data, client names or campaign successes.
- Test budget, channels and duration are determined once the real traffic volume and commercial risk limits have been reviewed.
- Cookie, explicit consent and personal data requirements are not assumed without legal approval; the measurement design is constrained by those decisions.