CPA, eCPA and EPC: calculate the economics before scaling
A worked hypothetical example connects submitted leads, approval, payout and break-even traffic costs.

Start with the event that earns commission
CPA can describe a commission model or a cost metric. Write down both the affiliate payout per payable action and your effective acquisition cost. Here, eCPA means traffic spend divided by payable actions; it does not mean the advertised commission. EPC means affiliate commission revenue divided by tracked clicks.
For cash-on-delivery offers, call-center approval and buyout are different events. Approval may mean an order is confirmed; buyout means the customer receives and pays for the goods. The contract determines which event earns commission, which reversals apply and when it becomes payable. Do not count every submitted form as revenue.
Work through one hypothetical cohort
Hypothetical example, not any network's results: spend $240 on 100 clicks, receive 20 submitted leads, and finish with 10 approved leads. Assume the contract pays $30 for each approved lead and those approvals remain payable. Commission revenue is 10 × $30 = $300; contribution after traffic spend is $60, before other costs.
Approval is 10 ÷ 20 = 50%. Cost per submitted lead is $240 ÷ 20 = $12, whereas CPC is $240 ÷ 100 = $2.40. The eCPA per approved lead is $240 ÷ 10 = $24. EPC is $300 ÷ 100 = $3. Traffic-only return on spend is $60 ÷ $240 = 25%.
Calculate the break-even point
At a $30 payout, $240 of traffic needs eight payable approvals to break even before other costs. With 20 submitted leads, that is a 40% approval rate. At the observed 10% click-to-payable-action rate, break-even CPC is $30 × 10% = $3. These thresholds change if quality or payout changes.
If this offer instead paid only on buyout, the ten approvals would not establish $300 revenue. Recalculate with the eventual payable buyouts. Subtract tracking, creative, payment and other relevant costs before describing net profit; allow headroom for reversals and sampling uncertainty rather than bidding exactly at the theoretical limit.
Compare complete cohorts and consistent definitions
Use the same click cohort, country, currency and attribution window for spend and revenue. Wait for the normal approval or buyout cycle. Show pending, rejected and reversed actions separately. Check whether a dashboard includes pending earnings, unique clicks or all clicks before comparing its EPC with your own calculation.
CashFactories, Everad, AdCombo and dr.cash are related CPA catalog entries to investigate, not evidence for the hypothetical rates above. Request current offer-specific terms and reconcile a bounded test. A larger headline payout alone does not establish stronger economics.