Three recruitment creatives, compared
The source report showed spend, CTR, CPC, CPA, impressions, clicks and conversions for each file. We transcribed the three fully visible rows below and checked the displayed rates against the raw counts.
| Creative | Impressions | Clicks | CTR | CPC | Conversions | CPA | Spend |
|---|---|---|---|---|---|---|---|
| Creative A | 172,532 | 1,232 | 0.71% | 0.19 | 218 | 1.08 | 235.26 |
| Creative B | 113,322 | 975 | 0.86% | 0.19 | 138 | 1.35 | 186.17 |
| Creative C | 53,503 | 738 | 1.38% | 0.21 | 94 | 1.68 | 158.31 |
Creative A
Lowest observed CPA and the largest conversion count.
Creative B
The same displayed CPC as Creative A, but a higher CPA.
Creative C
Highest CTR of the three, but also the highest CPA.
The source report did not identify the currency, attribution window, audience settings or conversion event. Cost figures are reproduced as report units rather than labeled with a currency symbol. Results describe this campaign only.
What was actually compared
The visible files were variants of the same Spanish dishwasher recruitment offer. The ads communicated a full-time Monday-to-Friday schedule, working hours from 9:00 AM to 4:00 PM, included meals and a monthly pay range. That makes the comparison more useful than placing unrelated campaigns side by side.
The campaign report contained a larger total of 512,362 impressions and 755 conversions, but that total included rows not fully visible in the supplied evidence. We excluded those unseen rows instead of assigning their results to these three creatives.
The highest CTR did not produce the lowest CPA
Creative C recorded the highest CTR at 1.38%, almost twice Creative A's 0.71%. If the review stopped at clicks, Creative C would appear to be the winner. Its CPA, however, was the highest of the three at 1.68 report units.
Creative A attracted clicks at a lower rate but recorded the lowest CPA at 1.08 and the largest conversion count at 218. In this report, the creative that generated the most economical recorded conversion was not the one that attracted the greatest share of clicks.
CPC alone did not explain the difference
Creative A and Creative B both displayed a CPC of 0.19, yet their CPAs were 1.08 and 1.35. Creative C's CPC increased only slightly to 0.21 while its CPA rose to 1.68.
That gap matters to a media buyer. A cheap click can still be a poor business result if the people clicking are less likely to complete the measured action. CTR and CPC describe the journey into the landing experience; CPA reflects what happened further down the funnel.
Delivery was not evenly distributed
Creative A received 172,532 impressions, while Creative C received 53,503. The platform may have allocated more delivery to the stronger result, or the files may have entered the campaign at different times. The screenshot does not tell us which explanation is correct.
For that reason, we treat the figures as observed campaign results rather than a controlled experiment. A cleaner future test would launch variants together, use the same audience and optimization event, and record the date range and attribution setting.
What a creative team should do with this result
The lesson is modest but useful: make several clear variants, let real delivery data challenge visual preference, and judge the result against the action the campaign was built to produce.
- Choose the primary metric before launch. Use CPA or qualified applications when the objective is recruitment, not CTR alone.
- Keep the job offer stable across variants so the visual treatment is the main meaningful change.
- Record audience, placement, dates, attribution window and conversion definition alongside every export.
- Review the application step when CTR improves but CPA becomes worse.
- Preserve losing variants. They may reveal a useful hook even when the full creative is not the winner.
About the evidence and this analysis
The creative and performance screenshot were supplied by a working editor and have been anonymized for this analysis. PosterSpeak AI did not operate the ad account and did not independently verify the platform's attribution.
AI assisted with organizing the supplied figures and drafting the page. The final calculations, claims, limitations and wording were reviewed against the original screenshot before publication.
