CTR — interest in the creative
CTR = clicks ÷ impressions. It shows whether the creative stops people. A high CTR alone means nothing: clickbait brings lots of clicks and few orders. Treat it as a first filter — "do people stop at all?".
CPC — cost per click
CPC = spend ÷ clicks. It depends on the auction, the GEO and the creative itself. A cheap click is only good if a conversion follows.
CR — landing page conversion
CR = leads ÷ clicks. If clicks are cheap and CR is low, the problem is almost always a gap between the creative's promise and what the person sees on the page.
EPC — what one click earns
EPC = confirmed revenue ÷ clicks. The key number to compare with CPC: if EPC is higher than CPC, the setup makes money. EPC already "contains" the conversion rate, the approval rate and the payout.
The one-line rule: EPC > CPC — scale; EPC < CPC — find the leak.
ROI — the bottom line
ROI = (revenue − spend) ÷ spend. Spent $500, earned $700 — ROI is 40%. It answers "was it worth it", but it lags: it's based on confirmed revenue, so it looks low in the first days.
Example: reading the numbers together
| Creative | CTR | CPC | CR | EPC | Decision |
|---|---|---|---|---|---|
| A | 2.4% | $0.18 | 6% | $0.31 | Scale |
| B | 4.1% | $0.11 | 1.2% | $0.07 | Change the landing page or switch off |
| C | 0.9% | $0.42 | 5% | $0.26 | Switch off: the click costs more than EPC |
Creative B has the best CTR and the cheapest click, yet it's the only one that earns almost nothing. That's exactly why decisions are made on EPC and ROI, not CTR.