Response Measurement
Part of Creative testing programmes
Reviewing results beyond the lowest acquisition cost
Review a creative test using conversion quality, value, timing and uncertainty alongside acquisition cost.
The lowest acquisition cost is useful only when both versions count the same kind of acquisition and those acquisitions are worth pursuing. Review the outcome chosen before launch, then check conversion quality, value, delivery and uncertainty before moving spend to a creative.
Check what the cost figure counts
Cost per acquisition divides spend by counted acquisitions. Ask which conversion action sits in the denominator: a form submission, a qualified enquiry or a purchase? Confirm both arms used the same definition and tracking setup. A creative that generates inexpensive enquiries may be the weaker choice if fewer become suitable customers.
For a purchase campaign, compare order value and, where the business can calculate it reliably, margin after relevant variable costs. For a lead campaign, use a consistently defined qualification stage and allow time for leads to progress.
Do not assign a precise lifetime value to a new lead without a defensible method.
Key Metrics to Review in Creative Testing (Australia)
- Conversion Action Type
- Form submission, qualified enquiry, purchase
- Order Value
- Track for purchase campaigns
- Margin After Variable Costs
- Use where reliable for lead-to-sale conversion analysis
- Lead Qualification Stage
- Consistently defined and time-allowed for progression
Read the full decision sheet
| Check | Question for the team |
|---|---|
| Delivery | Did both versions reach the intended audience and receive enough exposure to interpret? |
| Primary outcome | What happened on the measure chosen before launch? |
| Cost and value | What did each counted acquisition cost, and what value can be observed reliably? |
| Quality | Did enquiries, purchases or later customer outcomes differ in a way that matters? |
| Timing | Have conversions had a reasonable chance to arrive? |
| Uncertainty | Is the observed difference clear enough for the proposed decision? |
Use clicks and engagement to diagnose how the ad attracted attention, not as substitutes for the chosen outcome. Check the landing experience if a version draws interest without the expected action. Treat a small gap as inconclusive when the test cannot distinguish it from ordinary variation.
Compare both arms over the same sufficiently mature period and record when the data were extracted.
Keep attribution and lift separate
An attributed purchase is a purchase credited under a platform's conversion settings. It is not automatically an extra purchase caused by the ad. Do not label a lower attributed CPA as incremental growth without an appropriate measurement design.
Adopt a version when it improves the preselected outcome without a material decline in quality or another guardrail, and the comparison is reliable enough for the spend at stake. If cost improves while customer quality weakens, investigate the promise and destination before scaling. Record the trade-off rather than declaring a winner from one column.
