Your starting point
Prerequisites: Use the ideas and vocabulary from lesson 33; revisit its worked examples if needed.
Definition: Click-through rate is clicks divided by impressions. Conversion-per-click rate is conversions divided by clicks under a stated rule. Cost per conversion is spend divided by recorded conversions. Attribution assigns credit according to a measurement convention, not necessarily a causal mechanism.
Define the funnel stages
Impressions, clicks and conversions represent different opportunities and actions. Confirm that counts refer to the same campaign scope and period. Some systems count repeated actions or use attribution windows that extend beyond the display period. The academy sample uses nested synthetic counts for learning; a real export needs its own dictionary and validation.
Compare more than one metric
A campaign can have a high conversion rate but little reach, or many conversions at a high cost. Choose metrics that match the decision and acknowledge trade-offs. Avoid averaging percentages across campaigns with unequal denominators when you intend an overall rate. Pool the relevant counts, and report zero-conversion campaigns separately when calculating cost ratios.
Separate assigned credit from causal effect
A conversion credited to a campaign might have happened without it. Audience selection, returning customers and other channels can affect the outcome. Observational efficiency comparisons can guide a pilot or further investigation, but they do not isolate incremental impact. AI-generated recommendations should be checked for causal claims and missing cost or quality information.
Two worked examples
Click-through rate
A fictional campaign has 2,000 impressions and 80 clicks. CTR is 80 ÷ 2,000 × 100 = 4%. If it produces eight recorded conversions, conversion-per-click is 10%. The denominators differ, so the two percentages should not be labeled simply as the campaign’s rate.
Cost trade-off
Campaign A spends 100 for ten conversions, giving cost per conversion 10. Campaign B spends 90 for six, giving 15. A has a lower recorded cost ratio, but conversion quality, audience differences and attribution are unknown. The comparison is descriptive, not proof that shifting all spending to A will reproduce the result.
Where this becomes useful
A small organization compares two outreach messages. A transparent report can show counts, rates and cost while recommending a limited follow-up test rather than claiming one message universally works better.
Make an attempt
Use the complete workbench with Marketing campaigns. Create CTR and conversion-per-click fields, group by audience, and inspect pooled counts. Identify a recommendation that the data supports and one causal statement it does not.
Misconception check
Claim: Attribution credit is the same as causal credit.
Inspect the correction
Attribution follows a reporting rule. Determining what would have happened without the campaign requires a stronger design.
Independent application
Compare two synthetic campaigns using counts, two rates and cost per conversion. State attribution and zero-denominator rules, describe a trade-off, and propose a small verification step before generalizing the observed advantage.
Completion records practice and self-review. It is not certification of statistical competence.
