<b>Affiliate EPC: the metric everyone quotes, almost nobody understands</b>

<b>Affiliate EPC: the metric everyone quotes, almost nobody understands</b>

Thesis: earnings-per-click is reported as a stable average, but the underlying distribution is so heavy-tailed that the average barely describes any real campaign.

Context: network-published EPC figures are typically a 7- or 30-day mean across all affiliates promoting an offer. Analyses of affiliate-network data show conversion events clustering on a minority of traffic sources, with most clicks contributing near-zero.

Findings: a handful of high-intent placements generate the bulk of revenue; the modal click earns nothing. This means a network's advertised EPC is dominated by a few large affiliates and rarely predicts what a new promoter will see — often by an order of magnitude.

Caveats: EPC windows differ between networks, refunds and reversals are frequently excluded from the headline figure, and 'network EPC' blends incompatible traffic types (email, paid, organic) that have wildly different economics.

Implications: ignore network EPC for forecasting; compute your own per-source EPC after a full refund window closes.

What we still don't know: networks rarely disclose reversal-adjusted EPC, so the gap between gross and net EPC across the industry is essentially unmeasured.