Lesson 3 of 7 · 8 min read · intermediate
Native: Taboola, Outbrain, MGID and Revcontent
Native widgets sell clicks from news readers. See how native-to-search works, why site-level control matters, and what each network is known for.
Scroll to the bottom of almost any news article and you will find a grid of headlines with small pictures, often under a label such as 'Sponsored' or 'You may like'. That grid is a content recommendation widget. The ads inside it are native advertising: paid links designed to look like the reading material around them. Buying those clicks and sending them to a feed page is Native-to-search.
A newspaper sells the small boxes at the foot of each page to a broker. The broker fills them with teasers for other stories and takes payment whenever a reader follows one. The newspaper gets rent, the broker gets a cut, and the reader gets a suggestion that may or may not be worth their time.
Why native suits arbitrage
- The reader is already reading. Someone who has just finished an article is in the mood for another. An article-style content page is a natural next step.
- You usually bid per click. Native networks have long sold on a CPC basis, which makes the cost side of the sum easy to see, though conversion-based bidding is now common too.
- The audience skews towards news readers. That often means older visitors with interest in finance, health and home topics, which are the verticals advertisers pay most for.
- You can see where the click came from. Networks report by publisher site or placement, so poor sites can be removed.
The four names buyers mention most
| Network | What to know | Commonly cited strength | Commonly cited weakness |
|---|---|---|---|
| Taboola | Launched its Realize platform in February 2025, extending beyond bottom-of-article widgets into display, apps and device placements | Large premium publisher footprint and mature conversion bidding | Stricter review and higher entry cost than smaller networks |
| Outbrain (now Teads) | Outbrain completed its acquisition of Teads on 3 February 2025 and the combined company trades as Teads | Premium publishers and a tight content policy | Narrower tolerance for aggressive headlines |
| MGID | An independent native network with a wide long-tail publisher base | Low minimums and broad geographic spread | Quality differs a lot from site to site |
| Revcontent | A US-centred native network | Accessible self-serve buying | Smaller scale than the two leaders |
The work is in the site list
On Meta, one company owns all the inventory. On native, your ad may appear across thousands of independent sites, and they are not equal. A click from a national news site and a click from an obscure page crowded with widgets cost similar amounts but behave very differently once they reach your lander. Native buying is therefore largely an exercise in sorting.
- Tag every click with its sitePass the network's site or placement identifier into your tracker as a Sub ID so revenue can be matched to each source.
- Wait for enough dataA site with twenty clicks tells you nothing. Decide in advance how many clicks or how much spend a site gets before you judge it.
- Cut the losersAdd sites that spend without earning to a blocklist. Watch for sites where people click the native ad but almost never click anything afterwards: that pattern suggests accidental or low-quality clicks.
- Keep the winnersOnce you know which sites pay, some buyers move to an allowlist and bid more on those sites only.
- Automate the routineAutomation rules can pause a site when its cost passes a limit with no revenue, so the pruning does not depend on someone watching at 3 a.m.
Native policies in plain terms
Native networks review both the ad and the landing page, usually with human reviewers. Taboola's landing page policy, for example, requires pages with real content and substance rather than general search directories or indexes, bans countdown timers and false urgency, and requires that third-party ads on the page are clearly labelled as ads. Content that looks like a news story but is promotional must be marked as an advertorial.
Likely to pass review
- A headline that describes the article accurately
- A real article with useful information on the topic
- Related search terms that match the article
- Clear labels on ads and sponsored elements
Likely to be rejected
- A headline promising a secret, a giveaway or a shocking reveal
- A page that is only a list of search links
- Fake news-site styling, fake dates or timers
- Before-and-after or celebrity images used without basis
There is also a legal layer. Native ads imitate editorial content by design, so regulators expect them to be identifiable as advertising. In the United States the FTC Act treats an ad that misleads people about its commercial nature as deceptive, which is why native ad disclosure labels exist. The network adds the label on the widget. The buyer is responsible for honest labelling on their own page.
Key takeaways
- Native-to-search buys clicks from recommendation widgets on publisher sites and sends them to an article with a search feed.
- Inventory quality varies site by site, so site-level tracking, blocklists and allowlists are the core skill.
- Taboola launched Realize in February 2025; Outbrain completed its Teads acquisition on 3 February 2025 and now trades as Teads.
- Native networks review landing pages by hand and reject thin search-link pages, fake urgency and unlabelled advertorials.
- The buyer, not the network, answers to the feed for the quality of every placement.
Questions people ask
What is native to search arbitrage?
Native-to-search (N2S) arbitrage means buying clicks from native ad networks such as Taboola, Teads (Outbrain), MGID or Revcontent and sending those visitors to a page that carries a search feed. The buyer profits if the feed revenue per visitor exceeds the native click cost. It suits article-style landers because the visitor has just been reading a news site.
Is Taboola or Outbrain better for search arbitrage?
Neither is better in every case. Both have premium publisher networks and strict review. Buyers generally find results depend more on vertical, country, creative and how carefully sites are pruned than on the network name. Many operators test both with the same lander and keep whichever delivers better revenue per visit after blocking weak placements.
How do I stop bad sites wasting my native ad budget?
Pass the site or placement ID into your tracker, set a spend limit per site, and block any site that reaches the limit without earning enough. Look especially for sites with a high click rate on the ad and almost no activity afterwards. Automation rules can apply these cuts continuously so losses stay small.