Traffic arbitrage means earning the difference between what advertising costs and what the users it brings in generate. A media buyer purchases impressions and clicks in ad networks, sends that traffic to an advertiser’s product and gets paid for the users’ target actions. If payouts exceed the ad spend, the difference is their profit. Below we break down how the model works on real numbers, whether it is legal, and where to start if you have no budget of your own.
How arbitrage works: from click to paying user
The whole mechanic fits into one chain: spend, impressions, clicks, registrations, paying users, revenue. The advertiser, whether a dating product, an app or a service, is willing to pay for users because it knows their value, the LTV. The buyer takes on the job of bringing those users in, along with the entire risk that the advertising turns out to cost more than the payouts.
An important nuance: the advertiser does not pay out of altruism, but because the average paying user of a dating product generates more over their lifetime than the payout to the buyer costs. Both sides earn, and that is what has kept the model stable for more than fifteen years.
Here are the numbers of a typical month for a buyer on a PPS model (payment for the first purchase) at a $50 payout:
- Spend: $10,000 on Meta campaigns;
- Impressions: around 500,000;
- Clicks: around 15,000 visits to the landing page;
- Registrations: around 3,500 (click-to-registration conversion around 23%);
- Paying users: 3,500 × 8% ≈ 280;
- Revenue: 280 × $50 = $14,000;
- Operating costs: accounts, proxies, services, around $500;
- Profit: 14,000 − 10,000 − 500 = $3,500, ROI around 35%.
The combination of source, creative, landing page and offer that consistently delivers this result is called a campaign bundle, and hunting for those bundles is the essence of the profession. Getting to one working bundle usually takes a dozen losing tests, which is not a failure but the normal price of the search.
Notice how sensitive this math is to the conversion into a payer. If only 6% of those same 3,500 registrations started paying instead of 8%, that is 210 payers and $10,500 of revenue: the bundle goes from profitable to break-even at identical spend and identical lead cost. That is why professionals optimize toward the payment event rather than toward a cheap click or registration. They swap the pre-lander to match the creative’s audience, cut placements that produce empty registrations, and feed payment events back into the ad account through postbacks. The winner is not the one who guessed right the first time, but the one who tests faster and counts more honestly than the market, and teams with proper analytics systematically beat solo players who only ever see the lead cost in the ad account.
In plain terms: an analogy
Picture a wholesaler who buys fruit at a wholesale market for $1 per kilo and sells it at retail for $1.40. His business is not fruit, it is the price gap and the ability to find a market where things are cheap and a spot where they sell dear. A media buyer does the same with human attention: buys impressions wholesale in an ad network and sells already-interested users to the advertiser at retail. The word “arbitrage” comes from finance, where it describes earning on the price difference of one asset across different markets. Here the asset is traffic, the flow of people who clicked an ad.
One thing separates arbitrage from classic marketing: skin in the game. A marketer at an agency spends the client’s budget and gets a flat fee regardless of results. An arbitrage team spends its own money and earns only when it has done the math better than the market. So this profession has no room for pretty reports instead of results: there is ROI, and it is either positive or it is not.
Is traffic arbitrage legal
Short answer: yes. Buying advertising and getting paid for results is affiliate (CPA) marketing, a legal model used by some of the largest brands in the world. By eMarketer’s estimate, affiliate marketing spend in the US passed $10 billion for the first time in 2024. No statute anywhere forbids paying less for advertising than you earn from the users it brings.
So where do searches like “traffic arbitrage scam” and “can you go to jail for arbitrage” come from? From two real problems in the niche.
The first is grey methods. Some players run traffic to prohibited offers, use cloaking and creatives with fake promises, and work around payment system rules. That is a choice made by specific people and teams, not a property of the profession, in the same way that the existence of crooked accountants does not make accounting illegal. The consequences of those methods are concrete too: bans, lost accounts, and in some jurisdictions criminal exposure.
The second is the info-product industry around the niche. Most “I got scammed in arbitrage” stories are not about arbitrage itself but about purchased courses with income guarantees, “ready-made campaign bundles” for $200, and teams with no contract that simply never paid the share.
Working in a white team removes both problems: there is a legal entity and a contract, official payouts on a fixed date, an open bonus formula, and offers only from vetted advertisers. For a specialist it is ordinary employment, with a probation period, a salary and taxes (tax rates for Ukrainian sole proprietors, the FOP simplified regime, in 2026; the statute itself is article 293 of the Tax Code of Ukraine), the kind you can show to a bank or a consulate. Most Ukrainian buyers are registered as FOP, the country’s simplified single-tax regime for individual entrepreneurs, which sets a flat tax rate plus a unified social contribution and a NACE-style activity code for advertising services. A solo player carries all of these risks alone, legal and financial alike, from payment processor blocks to advertiser claims about traffic quality.
Verticals: dating, gambling, nutra, e-commerce
A vertical is the type of product you send traffic to. It determines everything: the funnel, the payouts, the risks and how fast you learn. Beginners often pick a vertical by payout size, which is a mistake: a high gambling payout is worth nothing if the first conversions only show up a week later and the budget ran out yesterday.
| Vertical | What gets paid for | Feedback cycle | Characteristics |
|---|---|---|---|
| Dating | registrations, subscriptions, rebills | 1–3 days | steady demand year-round, predictable funnels |
| Gambling | first-time deposits (FTD), redeposits | days to weeks | high payouts, harsh regulation and bans |
| Nutra | product orders (COD, SS) | days | depends on call-centre approval rates, aggressive creatives |
| E-commerce | purchases, ROAS | immediate | thin margins, fierce competition in the auction |
ROIcamp has worked exclusively with dating since 2011, and that is a deliberate choice rather than a limitation. Specializing in one vertical builds things you cannot assemble while spreading yourself across five: years of accumulated knowledge about creatives and audiences, private payouts from advertisers, and a higher share of profitable tests. Dating also has no seasonality, and traffic quality shows within 1–3 days, which for a beginner means a hypothesis gets validated in a day rather than a month. Why this vertical and how its economics work is covered separately in the dating traffic guide.
Traffic sources
A source is where impressions get bought. Each one has its own audience, its own moderation rules and its own cost of error, so the choice of source shapes a buyer’s daily work just as much as the vertical does. The main ones in 2026:
- Meta (Facebook/Instagram) is the primary source of dating traffic, around 70% of market volume. The best combination of volume and quality, and also the highest demands on account infrastructure, which requires constant farming.
- Google (Search/UAC) delivers quality traffic with search intent. Since the Dating & Companionship certification launched in 2025 (Google’s certification for dating ads) the source has become cleaner, but more expensive to enter.
- TikTok offers fast, cheap reach, and demands a video creative pipeline plus tolerance for bans.
- Push networks serve impressions inside notifications: a low entry threshold (from $50–100 a day) and lenient moderation, but a lower-quality audience.
- Native means recommendation blocks on content sites: a stable source for scaling proven bundles, with slow feedback for tests.
The rule of the niche: knowing one or two sources deeply beats knowing five superficially. Every source has its own moderation, its own creative formats and its own ways of killing an account, so depth beats breadth. That is exactly why teams assign buyers to sources instead of having everyone run a bit of everything: one specializes in Meta, another in TikTok, and each accumulates expertise that no single person could hold across all platforms at once.
Payout models: CPL, PPS, RevShare
The payout model defines what exactly the advertiser pays for, and therefore the entire math of a campaign. Typical dating payouts for Tier-1 countries:
| Model | What is paid for | Typical payouts (Tier-1) |
|---|---|---|
| CPL SOI | registration without email confirmation | $1–4 |
| CPL DOI | registration with confirmation | $2–8, Nordics up to $10 |
| PPS | first payment/subscription | $25–50, premium up to $100 |
| RevShare | share of revenue from the user | 25–50% lifetime |
The logic is simple: the deeper the action in the funnel, the higher the payout and the more traffic quality matters. SOI gives fast feedback and suits testing; PPS and RevShare reward those who can bring in an audience that actually pays rather than cheap registrations.
Cash flow affects the choice too. Lead payouts arrive quickly, but first pass through a hold, a traffic quality review period, and the advertiser pays only for approved conversions. RevShare stretches income across months: the first cheque is small, but rebills from paying users keep trickling in long after the campaign is switched off. Teams that hold traffic quality for years get private payouts above public rates and shortened holds, another advantage unavailable to a newcomer off the street.
Roles in an arbitrage team
The stereotype of the lone arbitrage guy with a laptop stopped matching the market a long time ago. A modern team is a production line with a division of labour:
- Media buyer, the central role: launches campaigns, tests bundles, reads the numbers and decides what to scale. The profession, its salaries and a typical day are covered in detail in who is a media buyer.
- Farmer: prepares and warms up ad accounts and maintains the account infrastructure. The most common entry position for beginners.
- Creative producer: a pipeline of static, video and UGC creatives built to buyers’ briefs. Creative is the main variable behind a bundle’s success, so the flow has to be continuous.
- Analyst: cohorts, LTV, reconciliation with advertisers and honest bonus calculations. Analytics is what separates a systematic team from “let’s run it and see”.
- Team Lead: runs a team of 3–5 buyers on one source, covering budgets, test priorities and people’s growth.
- Tech: tracker, postbacks, domains and advertiser integrations, the infrastructure without which a team’s numbers cannot be reconciled at all.
Compensation in the niche is standard: base salary plus a share of profit (10–20% and up), and the share is usually the bulk of a buyer’s income. In strong teams the bonus formula is open and buyers can see their own numbers in the tracker. If an interview promises to explain the formula later, that is a red flag.
For a beginner this structure means one thing above all: you do not need to know everything at once. You enter a team through one role and grow inside it, and a farmer sees live campaign bundles in week one and typically grows into an independent buyer within 12–18 months.
Solo or team: how much money you need
The question “what does it cost to start” has two very different answers.
Solo. The realistic minimum is $2,000–3,000:
- test budgets, from $1,500;
- accounts, proxies, antidetect browser, $100–300 a month;
- tracker and spy services, $50–150 a month;
- and above all, a readiness to lose all of it with nothing to show: the first bundles are almost always in the red, and the first 2–3 months go on infrastructure and understanding moderation.
A solo player’s first profitable bundle realistically takes 4–8 months, if the budget survives that long. Most burn through it before they learn enough, and fully recouping the investment often takes a year or more.
Team. Entry costs $0. The budgets, farming department, creative department, tracking and analytics already exist; losing tests are paid for by the team, and a beginner earns a base salary from month one while learning on working funnels instead of their own money. The typical track: first test budgets after 3–6 months, independent bundles after 12–18.
The difference is not only money, but the probability of ever reaching a result. A solo beginner learns from their own mistakes, each of which costs real dollars: a burned budget, a banned account, a month in the red. A beginner in a team walks over the same rakes using other people’s experience, with people around who have already stepped on every one of them and live campaigns that show what “done right” looks like. Then there is psychology: when a losing test does not eat your salary, decisions get made with your head rather than in panic.
The conclusion is unpleasant for the dream of quick solo money, but honest: the independent route is faster only on paper. A team gives you a slower start on budgets and a several-fold higher chance of reaching profit at all, plus a salary the entire way.
Key terms
The minimum vocabulary, without which no arbitrage chat makes sense:
- Bundle (campaign bundle): the combination of source, creative, landing page and offer that generates profit.
- Offer: the advertiser’s proposition, meaning the product and the payout terms for target actions.
- Spend: advertising costs over a period.
- Creative: the ad asset, whether banner, video or ad copy.
- GEO: the country or region traffic is sent to; payouts and traffic costs depend on it.
- Lead: a user who completed a target action, most often a registration.
- Approval rate: the share of leads the advertiser confirmed and paid for.
- Hold: the traffic review period during which the advertiser withholds payment.
- Rebill: a repeat subscription charge, the basis of long-term income in dating.
- Tracker: the system recording clicks, conversions and spend per bundle (Keitaro, Binom, Voluum).
- Postback: the conversion notification sent back from the advertiser into the tracker.
- Antidetect: a browser for running multiple ad accounts without linking them together.
Advice for beginners: do not learn terms from a list, they stick on their own once you start practising. It is enough to understand the logic: a bundle brings leads, leads pass approval and hold, payers bring rebills.
Typical beginner mistakes
We see these five mistakes in every second newcomer, and each one costs money. The list is short, but behind every item are hundreds of genuinely burned budgets.
- Single source. One platform equals one ban away from zero. The right way: keep at least two sources running; inside a team this is handled by assigning buyers to different sources.
- Running without a tracker. “I will just check the ad account” is a path into the red, because the ad account cannot see what happens after the click. The right way: a tracker from day one, with every bundle as a separate flow with its own stats.
- Chasing the lead instead of the paying user. A cheap lead that never pays is pure cost, and advertisers cut payouts fast for poor traffic. The right way: optimize toward the payment event and watch the conversion into a payer for each creative.
- Scaling before statistical significance. A profit on 20 leads is noise, not signal, and tripled budgets eat that kind of “profit” in a day. The right way: decide after 50–100 conversions per bundle and scale in steps of 20–30% of budget.
- Grey teams with no contract. An unpaid share is the most common bad story in the niche. The right way: a contract, a white structure and an open bonus formula, all before your first working day.
Where to start: learning and entry through a team
The main conclusion of this article compresses into one paragraph. Traffic arbitrage is a legal profession with transparent math, but in 2026 the barrier to entry is measured not only in money but in infrastructure: accounts, creatives, analytics, experience. Solo, you pay that barrier yourself, in cash and in months of losing tests. Through a team, it disappears entirely.
So if arbitrage looks like your story after reading this, do not start by buying a course. Free fundamentals plus practice inside a team beat any paid mentorship package. And if you have no experience at all, read how to become a media buyer with no experience, which walks step by step through entry via the farmer and assistant roles.
ROIcamp is a team that has worked with dating traffic and nothing else since 2011: 40 plus people, fully remote, white payouts and an open bonus formula. Entry roles with no experience required and vacancies for experienced buyers are all on the careers page.