If you cannot pin down how valid conversions are counted, how much rejections reduce revenue, and how far the settlement rate falls before cash is received, even the right method and execution may fail to produce real profit.
People in affiliate marketing often blame low earnings on insufficient traffic or lack of experience. But when you compare the dashboard with the settlement statement, another picture often appears: the traffic is fine, the method is the same, but the accounting is unclear. For the same offer, the revenue per click can differ from one operator to another, and the gap usually comes down to three numbers.
Number one: what counts as a valid conversion
Conversion reports in the dashboard often do not match settlement statements because the measurement rules are different. Clicks, registrations, first deposits, and retention can all be called conversions under different offers, while their settlement rates may differ by several times.
Three things need to be confirmed first: which event the advertiser actually pays for—registration or first deposit, order placement or confirmed receipt; whether there is a confirmation period between a reported conversion and a confirmed conversion; and who bears the cost of conversions rejected during that period.
In most networks, conversions first appear as pending and become confirmed only later. Pending figures can still be reversed at any time. If you use pending conversions to calculate earnings per click, you are recording uncertain money as revenue too early, which distorts later media-buying decisions. Confirm the rule with your account contact and write it in one sentence in your spreadsheet; that is much easier than reconciling the numbers afterward.
Number two: real conversion rate and rejection rate
The real conversion rate is not the dashboard conversion count divided by clicks. It is confirmed valid conversions divided by valid clicks. Two details matter: use confirmed conversions in the numerator, and remove obviously duplicated or abnormal clicks from the denominator.
The rejection rate is simple: rejected conversions divided by all reported conversions. A high rate usually comes from a few causes: duplicate submissions by the same user, ineligible information, orders that are later canceled or refunded, or traffic that fails the advertiser's quality review. These deductions are invisible on the day of the order and only become clear at settlement.
The importance of the rejection rate is that its effect accumulates. A channel may look strong on reported conversions, but if rejections are high, the audience quality is poor and scaling only makes the loss more consistent. Conversely, a smaller channel may show unimpressive conversion volume but a high confirmation rate and therefore better cash receipts. Channel-level accounting should use confirmed numbers, not nominal reported numbers.
Number three: the gap between the settlement rate and cash received
There is still a distance between the nominal commission and the amount actually received. That gap has three parts.
First are platform and advertiser deductions. Some are percentage-based, some use tiered rules, and some offers quote a gross rate, so the net rate has to be confirmed separately.
Second are invalid-traffic deductions. Duplicate clicks, proxy sources, and abnormal devices may cause part of the conversions to be removed, often without detailed line items.
Third is cash tied up by the settlement cycle. Monthly or bi-monthly settlement, minimum payout thresholds, and pending confirmation periods can leave money visible in the account but unavailable in cash. With the same absolute profit, a slower-paying channel puts much more pressure on cash flow.
Once all three parts are included, the accounting unit should become the actual cash received per thousand clicks. That is the number to compare with traffic cost, and it is more reliable than the nominal commission rate.
How the three numbers explain low earnings together
Earnings can be approximated as: clicks × real conversion rate × average confirmed rate × (1 − deduction ratio), then adjusted for the time cost of delayed payment.
A common reason earnings stay low even when the method is sound is that the assumptions are optimistic three times in a row: counting registration instead of first deposit, reported conversions instead of confirmed conversions, and gross rates instead of net rates. If all three are optimistic, the spreadsheet can show a profit while the actual cash result is a loss. The reverse also happens: some operators make all three assumptions too pessimistic and shut down channels that could have worked.
List these three numbers separately for each channel and the decision becomes much clearer. A channel with a high rejection rate and low confirmation rate produces ineffective scale no matter how much volume it has. A channel with a high confirmation rate but a slow settlement cycle should first be evaluated for cash-flow capacity before it is scaled.
One cost that is easy to overlook
When multiple platforms and accounts are running at the same time, reconciling the data becomes a cost by itself. If several accounts are repeatedly logged into from the same browser environment, operating traces can become linked, and it also becomes difficult to remember which account received which traffic and generated which commission. Reconciliation then depends on memory.
Giving each platform and vertical a fixed, independent login environment makes it possible to map account, channel, and data one-to-one, reducing the most error-prone reconciliation work. When account-level environment isolation and grouping are needed, an environment management tool such as PurpleMark can handle this part of the workflow.
Conclusion
Low earnings are rarely caused by one isolated issue. First define valid conversions clearly, then incorporate the rejection rate into the conversion calculation, and finally translate platform cuts, deductions, and settlement timing into the rate actually received. Once these three numbers line up, the figures in the report and the money in your account describe the same result.
