A native ad is a paid ad made to look and behave like the content around it: a sponsored post in a social feed, a sponsored product in search results, or a headline in a "recommended for you" block under a news article. Native ads blend in by format, and they still have to be labelled as ads.
For direct-response sellers, "native" usually means the recommendation networks such as Taboola and Outbrain, where a headline and a thumbnail on a publisher's site send the reader to an advertorial, and the advertorial sends them to the offer.
I'd judge a native campaign by sales per publisher site. The click price tells you very little.
Native advertising has two meanings, and that trips people up. To a brand marketer it means a sponsored article on a news site, written to sit next to the real ones. To a media buyer it means buying clicks from those blocks of headlines under articles, by the thousands, across hundreds of sites at once. Both are native ads. They're bought, built and measured very differently, and this post covers both, with more time on the second because that's where most paid-traffic money goes.
What a native ad is
An ad is native when it takes the form of the place it appears in. The IAB, the U.S. advertising industry body, puts it in terms of the ad being cohesive with the page content, the design and the way the platform behaves, so the viewer feels it belongs there. A banner sits in a box that looks like an ad slot. A native ad looks like the next post in the feed, the next result in the list or the next article to read.
What it can't do is pretend to be something other than an ad. Every native format carries a label ("Sponsored", "Ad", "Paid content"), and in the U.S. the FTC expects that label to tell a reader what they're looking at before they click (more on that below).
The three types of native ads
The IAB's Native Advertising Playbook 2.0 (May 2019) sorts native ads into three core types. Paid search ads technically fit its definition too, but the industry treats search as its own channel, so the playbook took it off the list.
| Type | Where you see it | Who usually buys it |
|---|---|---|
| In-feed / in-content | Social feeds, news feeds, sponsored products in a store's listings | Almost everyone who runs social ads |
| Content recommendation | "Recommended for you" blocks below or beside articles on publisher sites (Taboola, Outbrain and others) | Direct-response advertisers, affiliates, publishers buying traffic |
| Branded / native content | A sponsored article published by the site in the same format as its own stories | Brands with a content budget |
The playbook's 2013 edition had six types. The 2019 update folded promoted listings into in-feed (they're in-feed ads on commerce sites) and added branded content as its own type (IAB Native Advertising Playbook 2.0).
If you sell through paid traffic, the second row is the one to learn. A click on a content recommendation ad, in the playbook's words, takes the reader to an external page or to another page on the publisher's site. For you it's an external page, which means the landing page is yours to build, test and measure.
Native ads vs display ads vs social ads
The labels overlap, because almost every social ad is native by format. The useful difference is where the ad lives and how you pay for it.
| Native (recommendation networks) | Display | Social | |
|---|---|---|---|
| Looks like | A headline and a thumbnail among recommended articles | A banner in a fixed ad slot | A post in the feed |
| Where it runs | Many publisher sites through one network | Publisher sites and apps through ad exchanges | One platform's own feed |
| Usual pricing | Cost per click, with a bid | Cost per thousand impressions | Mostly auction on impressions, optimized for an action |
| Reader's mood | Reading, ready to read more | Reading something else | Scrolling |
| Typical landing page | An advertorial or presell | A product or landing page | Anything from a presell to a checkout |
The row I'd read first is the reader's mood. Someone who clicks a headline under a news story came to read, and they expect an article on the other side. Send them straight to a hard sales page and most of them leave. That one fact explains why native funnels look the way they do.
Native ad examples
The examples below are made up, for a fictional sleep supplement, to show the shape of each format.
- A recommendation widget headline. "Night-shift nurses swear by this 10-minute wind-down routine", with a photo of a dim bedroom, labelled "Sponsored" in a block of six headlines under a news article. The click goes to an advertorial on the advertiser's domain.
- An in-feed social post. A short video of the same routine in an Instagram feed, labelled "Sponsored", with a "Learn more" button to the same advertorial or straight to the sales page.
- A sponsored product listing. The supplement shown among a retailer's search results for "sleep aid", marked as sponsored, priced and ready to add to the cart.
- A branded article. "The science of the 3 a.m. wake-up", published on a health site under its own design with a "Paid content" line and the brand name at the top, written with or by the brand.
Same product, four native ads, and only the first is what an affiliate means when they say they "run native".
How native ads work for direct response
A native campaign on a recommendation network is a funnel with an extra step most other channels skip. The ad sells the click, the advertorial sells the idea, and the sales page sells the product.
- The ad. A headline and an image, written to make a reader curious enough to click. On these networks you usually pay per click, so you bid for clicks and set a daily budget.
- The presell. An advertorial, listicle or quiz on your domain that tells the story, frames the problem and points to the offer. This is the page that matches the reader's expectation of an article.
- The offer. The sales page or VSL (video sales letter), then the checkout. For an affiliate, this is often the vendor's page.
- The measurement. Which sites, headlines and images produced sales, and which only produced clicks.
The fourth step is where native campaigns are won. One campaign runs on hundreds of publisher sites at once, and they don't send the same readers. Taboola, for example, can write the publisher site, the site ID, the headline, the image and the device into your landing URL through tracking macros such as {site}, {site_id}, {title} and {platform} (Taboola help center). Put those into your UTM parameters, tie each sale back to them, and you can see which sites pay for themselves.
Then you block the sites that send clicks and no sales, raise bids on the ones that sell, and send the network a conversion for every sale so its bidding learns from buyers rather than clickers. Picture a campaign where the average cost per sale looks too high to keep: often a handful of sites are paying for the rest, and the campaign is profitable once the others are blocked. You only find that out if each sale carries the site it came from.
How much do native ads cost?
On Taboola, Outbrain and similar networks you mostly pay per click and choose the bid. What a click costs depends on the country, the device, the category, the sites you allow and how many other advertisers want the same readers, so any single number you read online is someone else's campaign. Sponsored articles on one publisher are a different purchase, usually a fixed price per piece agreed with the site.
The number to manage is cost per sale by site, which is the click price divided by how many of those readers buy. A cheap click from a site whose readers never buy is the most expensive click in the account. I'd rather pay more per click on sites that send buyers, and that only becomes visible once sales are tracked back to the site.
Native ad disclosure rules
Native ads borrow the look of content, so this is where regulators look. The FTC's position is that an ad must not suggest or imply that it is anything other than an ad. Its guidance names "Ad", "Advertisement", "Paid Advertisement" and "Sponsored Advertising Content" as labels readers understand, placed in front of or above the headline (FTC native advertising guide).
- The ad unit carries the network's label. You don't control that one, but your headline mustn't contradict it, for example by posing as a news report.
- The landing page needs its own label when it looks like an article. An advertorial reached from a native ad is still an ad.
- The claims follow the same rules as any ad: real testimonials, and evidence behind health claims. What is an advertorial covers the page-level rules in more detail.
The networks review the landing page as well as the ad, and each has its own policy on top of the law. This is general information about U.S. rules, not legal advice.
Running native ads with ElasticFunnels
Native is the channel where the page after the click matters most, and where a sale with no site attached to it is close to useless for optimizing. Those are the two things ElasticFunnels is built around. Every feature below is on every plan.
- Build the presell. The AI page builder builds any page you describe, with tuned prompts for advertorials, presells and listicles, ready to edit.
- Test the angle on one URL. A same-URL split test runs up to 5 versions of the advertorial under the address the network already approved, so the campaign keeps running while you test.
- Send each sale back to the network. The Taboola integration sends a server-to-server conversion with the click ID, event name and revenue, and the Outbrain integration sends purchases and upsells server-side with the Outbrain click ID.
- See revenue by site. Pass the network's site macro in a UTM parameter and the UTM Performance report shows revenue and conversions per value, which is per publisher site.
- Change the page by source. Page events can show different content depending on a query parameter or the traffic source, so one advertorial can open with a different line for a different placement.
Affiliates will find the full presell-to-postback setup on the ElasticFunnels for affiliates page.
- Don't judge a native campaign in its first days on the blended number. Wait until each site has enough clicks to say something, then cut by site.
- Write the headline for the reader of the site it runs on. A curious headline that the advertorial doesn't pay off gets clicks and refunds.
- Never send cold native traffic straight to a checkout. These readers clicked to read.
- If a headline only works by looking like news, it's a liability for your account however well it clicks.
Native ads have a reputation for clickbait, and the bad end of the channel earns it. The good end is a well-labelled story that a curious reader chose to click, on a site whose readers actually buy, measured down to the sale. That's a lot of moving parts for one click, which is why I'd rather have the page, the test and the tracking in one place before I spend a dollar on it.
