The key idea
First separate the campaign objective and metric families. View-through conversions record a conversion following exposure without a recorded interaction; they do not alone establish incremental advertising impact.
The starting point
- Available exports by campaign, creative and placement, with dates and currency
- Impressions, clicks, cost, conversions and value, with column definitions
- Viewability, reach and frequency where available, plus targeting context
1. Define the question before aggregating
A reach campaign and a conversion campaign need different priorities. Separate prospecting and remarketing where possible. Preserve reporting limitations: dimensions may not be available together, and totals across exports may overlap.
2. Check metric definitions
Calculate CPM = cost / impressions × 1,000, CTR = clicks / impressions and CPC = cost / clicks. Show denominators and volumes. Understand the base of viewability rates before comparing them. Keep view-through conversions distinct and check their window.
3. Identify segments to investigate
Ask for combinations concentrating spend or showing change. Examine high repetition, declining creative performance or questionable placements without automatically concluding fatigue or fraud. Link every signal to data and a possible check.
4. Turn findings into learning
Design a creative, placement or frequency test using available controls. Define an expected outcome and spending limit. Estimating incrementality requires a suitable experiment; descriptive reporting does not replace a meaningful comparison group.
A practical example
Analyse the attached display exports. Check granularity, duplicates, dates and definitions. Separate reach, engagement and conversion. Calculate metrics from underlying counts and keep volumes visible. Distinguish view-through from other conversions. Produce sourced observations, hypotheses and missing checks, then three possible tests. Do not assign causality or conclude fraud from an isolated signal.
| Signal | Useful question | Avoid concluding |
|---|---|---|
| High frequency | Which audience and period? | Certain fatigue |
| Low CTR | Is clicking the format’s objective? | Definitely poor creative |
| View-through conversions | Which window and remarketing role? | Sales caused by advertising |
| Concentrated spend | Are placement and volume relevant? | Proven fraud |
How to assess progress
- Check compatible segments and export limitations.
- Do not add differently defined conversions without documented deduplication.
- Keep budget decisions separate from generated analysis.
Official sources
Google Ads · Conversion tracking dataFind the campaign hidden by the average
A synthetic CSV covers four campaigns across two periods. The question: does improving overall return hide a problem?
View detailsThese demonstrations use fictional briefs, documents and synthetic data. They are examples of method, not client results.
MindVerve
Want this workflow built for your team?
MindVerve can design your display review framework: exports, checks, segment analysis and a test plan. The deliverable supports decisions with your acquisition lead, with assumptions and limits visible.
Discuss my AI workflowFive areas. One coherent direction.