Traffic Arbitrage Metrics Explained: ROI, CPA, EPC and Break-Even CPC

Last updated: 2026-10-28

Short answer: Traffic arbitrage metrics tell you whether you buy clicks for less than they earn. Track CPC, CR, CPA, EPC and ROI; the key comparison is EPC against CPC, and break-even CPC equals CR × payout × approval rate. Add hidden costs, payment holds and enough data before you scale a campaign.

Traffic arbitrage sounds simple: buy traffic for less than you earn from it. The whole game is in that "less than". And you can't know whether you're there without a handful of numbers.

Beginners usually look at one metric — clicks, or the shiny ROI number in a tracker — and make decisions on half the picture. This article gives you the full set of formulas, shows how they connect, and walks through an illustrative example so you can do the maths on a napkin (or better, in a spreadsheet) before you spend a single euro.

Risk note: Not financial advice. Paid traffic can lose money quickly, and past campaign results don't predict future ones. Always follow ad network and offer rules.

How does traffic arbitrage make money?

You pay a traffic source (an ad network, a social platform, a push or native network) for impressions or clicks. Those clicks go to an offer. When a visitor completes a target action — a registration, a lead form, a purchase — the advertiser or affiliate network pays you a payout. Profit = payouts − traffic cost − other costs.

Everything below is just that sentence broken into parts.

Which metrics matter in traffic arbitrage?

Traffic-side metrics

• Impressions — how many times your ad was shown.
• CTR (click-through rate) = clicks ÷ impressions.
• CPM (cost per mille) = cost ÷ impressions × 1000.
• CPC (cost per click) = cost ÷ clicks.

Offer-side metrics

• CR (conversion rate) = conversions ÷ clicks. (Be clear whether it's clicks on the ad or clicks to the offer page — trackers can show both.)
• Payout — what you earn per approved conversion.
• Approval rate — the share of conversions the advertiser actually accepts. Leads can be rejected for low quality, duplicates or fraud.

Money metrics

• CPA (cost per action) = cost ÷ conversions.
• Revenue = approved conversions × payout.
• EPC (earnings per click) = revenue ÷ clicks.
• Profit = revenue − cost.
• ROI = profit ÷ cost × 100%.

How they connect: the two numbers that matter most

If you remember only two comparisons, remember these:

1. EPC vs CPC. If you earn more per click than you pay per click, the campaign is profitable on traffic cost. If EPC < CPC, you lose on every click.
2. CPA vs payout. If one conversion costs you less than it pays, you're in the green.

They're the same truth seen from two angles. Use whichever is easier with the data you have.

How do you calculate break-even CPC?

This is the most useful planning formula:

Break-even CPC = CR × payout × approval rate

It tells you the maximum you can pay per click before you start losing money — before you launch anything.

Illustrative worked example

All numbers below are invented to show the maths. They're not benchmarks for any vertical, network or geo.

You're testing a lead-generation offer:

Step 1 — Break-even CPC: 0.04 × €8 × 0.75 = €0.24

Your CPC (€0.20) is below €0.24, so on paper there's room.

Step 2 — What €100 buys: €100 ÷ €0.20 = 500 clicks

Step 3 — Conversions: 500 × 4% = 20 leads → 20 × 75% = 15 approved

Step 4 — Revenue and profit: 15 × €8 = €120 revenue €120 − €100 = €20 profit

Step 5 — ROI, CPA, EPC:

• ROI = 20 ÷ 100 = 20%
• CPA (on all leads) = €100 ÷ 20 = €5; on approved leads = €100 ÷ 15 ≈ €6.67
• EPC = €120 ÷ 500 = €0.24

Now the sensitivity check. What if the real CR is 3% instead of 4%? 500 × 3% = 15 leads → ≈ 11 approved → ≈ €88 revenue → −€12, ROI ≈ −12%.

One percentage point of CR flipped the campaign from profit to loss. That's why you don't trust a single number, and why "small" changes in landing pages or targeting matter so much.

Which metrics do beginners forget?

Hidden costs

ROI in a tracker usually counts only traffic spend. Your real costs may include:

• tracker subscription,
• creatives (design, video, AI tools),
• landing page hosting and domains,
• payment fees and currency conversion.

Add them to a separate "overhead" line in your sheet and calculate net ROI monthly.

Holds and payment terms

Networks often pay on a schedule, and some conversions stay "on hold" until verified. Your cash flow can be negative even when the campaign shows profit. Know the payment terms of your network before scaling.

Statistical noise

Twenty conversions is not a lot. A campaign that looks great after 3 conversions can look terrible after 30. Set a minimum data threshold before judging (we cover test budgets and kill rules in a separate article).

Your arbitrage metrics sheet

Set up one row per campaign per day with these columns:

Formulas:

• Revenue = Approved × Payout
• Profit = Revenue − Cost
• ROI = Profit ÷ Cost
• EPC = Revenue ÷ Clicks
• CPA = Cost ÷ Conv.

Add one cell at the top with your break-even CPC for each offer, and colour any day where actual CPC exceeds it. That one cell saves more money than most "secret" strategies.

Stay on the right side of the rules

Arbitrage has a reputation problem because of misleading ads and banned verticals. Don't be part of it:

• Read the offer's allowed traffic sources and restrictions.
• Follow the advertising policies of the platform you buy from.
• Don't use fake claims, fake celebrity endorsements or deceptive landing pages.
• Avoid offers you'd be embarrassed to explain to a friend.

Rule-breaking campaigns get banned, payouts get clawed back, and your numbers become fiction.

Where to learn the full process

These formulas are the starting point of the Phoebus Lab Traffic arbitrage course on metrics, tests and funnels. It begins with writing your money loop (spend, conversion, payout, profit) and builds toward a funnel table with a spend cap and a closed test with a written conclusion, even a red one. Plan a separate test budget (the course suggests €300–500) and treat it as money you can lose. No profit is promised.

Related reading: How to Test Ad Creatives in Traffic Arbitrage: Test Budget, Kill Rules and a Log Sheet · The first week of bought traffic is an account, not a creative contest

— FARM