Separate rising impression costs, weaker link click-through rates and fewer attributed purchases. Use a worked example to choose the first check instead of guessing.
When cost per purchase rises, separate three questions: Did impressions become more expensive? Did fewer people click your link? Or did those clicks produce fewer attributed purchases? All three can raise acquisition costs, but they call for different checks. Start by locating the change rather than declaring the creative ineffective.
Make the periods comparable
Compare the same account and campaign across periods with similar weekday coverage. Keep currency, attribution settings and conversion definitions consistent. All clicks and link clicks are different measures; do not use them interchangeably. Purchases attributed to Meta are also different from all orders in your store. Recent clicks may still produce conversions that appear later, so do not treat the last few days as final results.
Break the change into three stages
Impressions: Did CPM increase?
CPM is the average cost of a thousand impressions. If link CTR and the purchase-to-click ratio stayed similar while CPM rose, start with delivery costs. Compare the same placements and look for changes in where spend went. A higher CPM alone does not prove that competition increased: the mix of placements or delivery can also change the average.
Clicks: Did link CTR fall?
Link CTR divides link clicks by impressions. With similar CPM, a lower link CTR can mean fewer link visits for the same budget. Inspect individual ads before judging the campaign average. Did one creative decline, or did a lower-CTR ad simply receive more impressions? These situations may require different next steps. Do not infer creative fatigue from frequency alone.
Purchases: Did the post-click result change?
If clicks and click costs stayed similar but attributed purchases fell, check product availability, prices, landing pages, checkout and measurement changes. Dividing attributed purchases by link clicks is a reporting diagnostic, not a verified person-by-person funnel conversion rate. The two counts need not represent exactly the same people or events.
Worked example: CPA rose by 50%
Consider two hypothetical periods with 10,000 impressions and 200 link clicks each. The first spends TRY 1,000 and records ten purchases. The second spends TRY 1,200 and records eight. Link CTR remains 2%. CPM rises from TRY 100 to TRY 120, while cost per purchase rises from TRY 100 to TRY 150. Cost per link click increases from TRY 5 to TRY 6, and purchases divided by link clicks decreases from 5% to 4%.
The example contains two signals: impressions became more expensive, and purchases per link click decreased. There is no link-CTR decline supporting a claim that the ad message stopped working. Compare placement-level CPM and check whether product or checkout conditions changed. A difference of two purchases is a small sample, so it should not determine a budget change on its own.
Turn the review into a practical next step
Record the campaign, dates, changed metric and one specific check. Set a review date and a success measure before making a test. For example, monitor whether link CTR improves without purchase costs deteriorating. The AdLens analytics overview explains where campaign reporting fits into that review. Missing measurements should remain missing; limited evidence should not become a confident claim about the cause.
For the Turkish version of this guide, see Facebook reklam analizi. Metric fields can be checked in Meta’s official Ads Insights SDK.
