Listed players
SST2.44▼ -7.58%TIG40.00▲ +3.90%TEAD0.56▲ +3.77%PERI8.50▼ -2.97%TBLA3.23▼ -2.71%INUV0.57▼ -1.74%AV10.06▼ -1.59%GOOGL343.50▲ +1.56%SNAP5.58▼ -1.24%PINS19.26▼ -1.03%MSFT517.53▲ +0.92%PPLI41.28▲ +0.81%IOS32.24▲ +0.44%META728.08▲ +0.30%GDDY97.21▲ +0.24%DV13.49▲ 0.00%MCHX1.29▲ 0.00%
Ticker byClearTrust

Lesson 1 of 7 · 8 min read · intermediate

This lesson counts towards the ClearTrust Media Buying & Compliance certificate. Enrol with your email to record your progress and scores.Get certified, free

Traffic sources compared: where arbitrage clicks come from

Meta, native, TikTok, Demand Gen, search and push all sell clicks, but they sell very different people in very different moods. Here is the map.

Every search arbitrage campaign starts with a purchase. The media buyer pays a traffic source for a visitor, sends that visitor to a page carrying a search feed, and hopes the feed pays more for the visitor than the visitor cost. The feed side is fairly fixed once a contract is signed. The buy side is where most of the daily work, and most of the daily risk, sits.

Think of a market stall that sells lemonade. The recipe barely changes. What changes every morning is where you stand: outside the gym, next to the school gate, by the office block. Each spot has different rent, different crowds and a different landlord with different house rules. Traffic sources are those spots.

The arbitrage loop

1/7
The gap$Arbitrageurbuys and sells clicks⌂Traffic sourcee.g. a social app◉Visitora real person▤Landerarticle + search terms⇆Search feedsponsored results★Advertiserpays per click$The spreadrevenue minus cost
1
Two prices for the same person

A click on a social app might cost $0.20. A click from the same person on a search ad might be worth $1.00 to an advertiser. Search arbitrage lives in that gap, like a trader who buys fruit at the farm gate and sells it in the city (all figures here are illustrative).

  1. Two prices for the same person: A click on a social app might cost $0.20. A click from the same person on a search ad might be worth $1.00 to an advertiser. Search arbitrage lives in that gap, like a trader who buys fruit at the farm gate and sells it in the city (all figures here are illustrative).
  2. Buy a visitor: The arbitrageur runs an ad on a traffic source and pays for each click. Say the cost per click is $0.20. That money is spent whether or not the visitor ever earns anything back.
  3. The visitor lands on a page: The person who clicked arrives on the arbitrageur’s landing page: usually a short article with a block of related search topics underneath.
  4. A search shows sponsored results: If a topic interests them, the visitor taps it and sees a results page. The ads on it come from a search feed: the same advertisers who bid on a big search engine, shown on this smaller site.
  5. The advertiser pays for a click: The visitor clicks one sponsored listing and goes to the advertiser’s site. The advertiser is charged, say, $1.00. This is the only moment in the loop when new money enters.
  6. A share comes back: The search engine and the feed provider keep their cuts and the arbitrageur receives, say, $0.56 for that ad click: the RPC. But only about half of visitors click an ad, so the average visitor earns $0.28 (RPV).
  7. What is left is the spread: $0.28 earned minus $0.20 paid leaves $0.08 per visitor: the spread, a 40% ROI. Across 10,000 visitors a day that is $800. If the ad price rises to $0.30 or fewer people click, the same loop loses money just as fast.

Two kinds of visitor: interrupted and asking

The single most useful distinction is whether the person was looking for something. On Meta, TikTok or a news site, people are scrolling. An ad interrupts them, so the creative has to create curiosity from nothing. On a search engine, the person typed a keyword and already has search intent. Interrupted traffic is plentiful and often cheap, but it must be turned into intent on the landing page. Search traffic arrives with intent, but there is less of it and it is priced accordingly.

Interrupted (push) traffic

  • Meta, TikTok, Snap, Pinterest, native widgets, Demand Gen
  • Huge volume, priced by auction on impressions
  • The ad must spark interest, so creative quality decides the cost
  • Visitor intent is weak until the lander focuses it

Asking (pull) traffic

  • Google Ads search, Microsoft Advertising search
  • Limited by how many people search each term
  • The keyword does the work, the ad only has to match it
  • Strong intent, but the buy price is closer to the sell price

The main sources side by side

A qualitative guide. Prices are deliberately left out: they move daily by country, vertical and season.
SourceModel nameTypical strengthsTypical weaknessesPolicy strictness
Meta (Facebook, Instagram)Social-to-searchLargest scale, strong automated delivery, feeds revenue signals back wellAccount bans, sensitive to edits, tight rules on finance and health claimsHigh
Taboola, Teads (Outbrain)Native-to-searchReaders already in article mode, site-level control, older audiencesQuality varies widely by site, manual review, needs blocklistsMedium to high
MGID, RevcontentNative-to-searchLower entry cost, wide long-tail inventoryMore variable traffic quality, some feeds will not accept itMedium
TikTok, Snap, PinterestSocial-to-searchYounger or niche audiences, video formats, less crowded auctionsShort creative life, weaker fit for high-value verticalsHigh (TikTok) to medium
NewsBreakNative-to-searchUS news readers, self-serve conversion biddingUS-centred, limited volume next to MetaMedium
Google Demand GenDisplay-to-searchYouTube, Discover and Gmail reach, strong bidding toolsGoogle Ads destination rules name arbitrage directlyVery high
Search ads (Google, Microsoft)Search-to-searchReal intent, keyword-level controlThin spread, strict destination rules, limited volumeVery high
Push and pop networksPush notification traffic, pop trafficVery cheap clicksHigh invalid-traffic risk, widely refused by feedsLow at the source, refused downstream

Why the feed cares where the click came from

The advertiser at the end of the chain pays for a click because they expect a real person with a real need. If the visitor was tricked, bored or automated, the advertiser's conversion rate collapses and they complain or opt out. Google's rules for Related Search on Content therefore reach back to the ad that started the journey: the traffic source creative must be relevant to, and accurately describe, what the user will see on the destination page, and since 1 November 2025 the exact ad text must be passed to Google in the referrerAdCreative parameter when the traffic comes from a source the publisher controls.

1 Nov 2025Date from which Google requires the referrerAdCreative parameter on related search for content pages when traffic arrives from a source under the publisher’s controlSource: Google AdSense Help: Related search for your content pages

How buyers actually choose

  1. Start from the feed, not the sourceThe feed contract lists approved sources, countries and verticals. That list is the menu. Anything off it is a breach, however profitable it looks.
  2. Match the audience to the verticalA vertical such as retirement insurance needs older readers, which points to Facebook and news-site native. A consumer gadget topic may suit short video.
  3. Check you can measure itA source is only usable if clicks can be tied to revenue through a tracker and a Click ID. Without that the buyer is guessing.
  4. Test small, on one sourceLearn one platform's auction and policies before adding a second. Splitting a small budget across five sources teaches nothing about any of them.
  5. Spread once it worksRelying on one ad platform is platform risk. A second source is insurance against an account ban, not just extra volume.

The rest of this track takes each source in turn. Keep one idea in mind throughout: the buy side and sell side are joined. What you say in the ad, who you show it to and where they land all travel downstream to the feed, and from there to the advertiser who funds the whole thing.

Key takeaways

  • Traffic sources split into interrupted traffic (social, native, Demand Gen) and intent traffic (search); the first is bigger, the second is closer to the sell price.
  • Every click must satisfy two rulebooks: the ad platform’s policy and the feed’s list of approved traffic sources.
  • Google’s RSOC rules require the ad to describe the destination accurately and, from 1 November 2025, the ad text to be passed in referrerAdCreative.
  • Cheap sources such as push and pop are widely refused by feeds because of invalid-traffic risk.
  • Choose sources from the feed’s approved list, match audience to vertical, and diversify once one source works.

Questions people ask

What is the best traffic source for search arbitrage?

There is no single best source. Meta is the largest by scale and the most widely used for related-search feeds, native networks such as Taboola suit article-style landers and older audiences, and search-to-search offers real intent on a thin margin. The right choice depends on which sources your feed provider approves, your vertical and your country. Start with one approved source and learn it properly.

Can I use push or pop traffic for search arbitrage?

In most cases no. Push and pop clicks are cheap, but they carry a high share of accidental, low-intent and automated visits. Major feeds either refuse these sources outright or reduce payouts sharply when advertiser results are weak. Sending unapproved traffic to a feed is a contract breach that can end the account and trigger revenue deductions.

Do I need permission from the feed to buy traffic from a new source?

Yes. Feed agreements normally list the traffic sources, countries and sometimes the verticals you may use. Adding a new source without approval risks clawbacks and termination even if the traffic is genuine. Ask the feed provider in writing, and expect to share example ads and landers before they agree.

Next: Meta: buying arbitrage traffic on Facebook and Instagram