Working space A / Invented data
Read the sheet.
Question the story.
Three samples can support several observations. None can explain its own cause. Use the bench to separate arithmetic from interpretation.
The same definitions.
Three different samples.
- Impressions
- 10,000
- Clicks
- 400
- Spend · AUD
- $800.00
- Conversions
- 20
A starting point: 20 recorded conversions from 400 clicks. The numbers do not tell us whether those conversions were useful.
Definitions: Google Ads: clickthrough rate, average CPC and average CPA. Reviewed 19 October 2026.
All spend shown here is fictional AUD. These examples assume one consistent conversion definition; real reporting requires checking the selected scope and measurement settings.
Try the interpretation
Sample B has more clicks
than sample A. What follows?
Choose a conclusion, then examine the reasoning.
The questions
outside the sheet.
What counts?
Agree what the conversion represents. A recorded form completion and a useful enquiry need not be the same thing.
What changed?
Check the reporting period, measurement definition and relevant changes to the audience, offer or website.
What remains unknown?
These samples contain no revenue, margin, enquiry quality or controlled experiment. Do not infer profitability or a cause from them.
You have an observation. Now make the next question testable.
Draft a hypothesis