Affiliate marketing is not about creatives or luck — it is about math. Those who can model a funnel's unit economics before launch lose hundreds of dollars on tests; those who cannot lose thousands. Here are all the core affiliate metrics in plain language, with formulas and benchmarks.
Top-of-funnel metrics
- CTR (click-through rate) — clicks ÷ impressions × 100%. Shows whether the creative hooks. Benchmarks: push and native — 0.1–1%, Facebook — 1–3%, TikTok — 1–2%. Low CTR is fixed by changing the creative, not the bid.
- CPC (cost per click) — spend ÷ clicks. Means nothing on its own: a $2 Tier-1 click can pay back better than a $0.02 Tier-3 click.
- CPM — cost per thousand impressions. Useful for comparing auctions and placements.
Conversion metrics
- CR (conversion rate) — target actions ÷ clicks × 100%. Measure it per funnel step: lander CR, offer CR. A prelander converting at 3% versus 1.5% halves your lead cost on the same traffic.
- CPL / CPA — spend ÷ leads. The main cost metric you compare against the payout.
- Approval rate — the share of confirmed leads. Critical in nutra and COD e-commerce: a $5 CPL at 30% approval is really $16.6 per paid lead.
Money metrics
- EPC (earnings per click) — revenue ÷ clicks. The key metric for comparing offers: an $800-payout offer with $0.4 EPC is worse than a $300 offer with $0.9 EPC. Look at EPC for your traffic type, not the network average.
- ROI — (revenue − spend) ÷ spend × 100%. ROI of 100% means you doubled your money. Most stable funnels live at 20–80%; anything consistently above 100% does not last long.
- Hold — time until payout. Not an efficiency metric but a cash-flow one: a funnel at 60% ROI with a 30-day hold needs several times more working capital than one with a weekly hold.
Unit economics before launch
The forecast formula is simple: profit = clicks × (EPC − CPC). Before testing, compute your ceiling CPL: payout × expected approval. Example: $24 payout, 40% approval → $9.6 max CPL. If spy tools and network stats show leads in that geo never come cheaper than $7 for beginners — there is no margin, skip the funnel; you just saved the test budget before spending it.
When to kill a campaign
The sampling rule: judge a placement or creative after spending 2–3 conversion costs, not earlier. Spent $20 against a $10 target CPL with zero leads — cut it. Leads exist but ROI is −30%? Fix the weakest link by metrics first: low CTR — creative, low lander CR — landing page, low approval — call center or geo. Metrics exist precisely to show where the funnel leaks — a tracker shows this in detail, and we have a separate blog guide on setting one up.
Model before launching, cut by numbers instead of feelings — and your test budget becomes an investment rather than a lottery. Bring your funnel math questions to the Beginner Guides section of the forum.