Native advertising on mobile: you supply components, the publisher renders the advertisement, and the thumb decides the measurement

The thing that makes mobile native different from every other display format is that you never see the finished advertisement before it runs. You hand over components, a headline, a line of text, an image or a video, an advertiser name and a call to action, and the app or the site renders them into its own template at whatever size that device and that feed position allows. So the craft is not designing an advertisement. It is supplying parts that survive being truncated, cropped and stacked by a layout engine you do not control. Everything else about mobile native follows from that: how the copy is written, why the click rate is unreliable, and why the landing page is half of what you are paying for.

Where mobile native actually runs

Surface	How it is bought	What it is good at	The specific risk
Social in feed	Direct in each platform's own auction	The majority of native spend, with the best targeting and the best creative tooling	It looks like native and behaves like performance media, so it gets judged on last click and starved
In app native, inside other people's apps	Programmatically, through a demand side platform against app inventory	Reach at low cost per thousand, including audiences who barely use the mobile web	Accidental clicks, and app inventory quality that has to be checked rather than assumed
Content recommendation units on mobile web	Through recommendation platforms such as Taboola or Outbrain	Volume to a content destination, cheaply	The unit sits below an article, so the audience was finishing something else
Publisher app and news feeds	Direct with the publisher, or programmatically	Context and a readership that trusts the surrounding content	Priced for the context, so it has to be judged on brand outcomes rather than clicks

The three categories, and the four tests

The IAB's Native Advertising Playbook 2.0, published in May 2019 as an update to the original from December 2013, cut the earlier list of six native types down to three, folding promoted listings into the first of them.

• In feed and in content units, placed inside an article, a product listing or a social feed and styled to match it. This is where almost all mobile native sits.
• Content recommendation units, the grids of paid links beside or below an article, which usually send the reader to a different site.
• Branded or native content, paid material published in the publisher's own format.

The playbook then judges any unit on four criteria: design, location, ad behaviour and disclosure. Those four are a usable checklist when a seller offers you something they are calling native. If it does not match the design of its surroundings, does not sit in the content flow, does not behave like the content around it, or does not carry a clear label, it is a display advertisement with a nice name.

Writing for a template you cannot see

• Front load everything. Assume the headline is cut at roughly half of what you wrote and the body text at one line. Put the offer, the product and the reason in the first few words, then write the rest for the renderings that show more.
• Never put essential words inside the image. Crops differ by placement, and a square asset becomes a wide one somewhere. Text in an image is also the first thing a platform's own review process objects to.
• Design the image for a thumbnail. One subject, high contrast, readable at the size of a postage stamp, because that is the size it will be in a feed on a phone held at arm's length.
• Supply every aspect ratio the specification asks for. Missing assets get filled by automatic cropping, which is how a product ends up half out of frame.
• Make the advertiser name do work. On most mobile native units your brand name appears as a small line of text rather than a logo, so it has to be recognisable as plain words.
• Video needs to work silently. Autoplay is muted in feeds, so the first two seconds have to carry the idea without sound and without a voiceover being missed.

The thumb problem, and what to measure instead of a click

A phone screen is small, native units sit in the scroll path, and fingers are imprecise. That produces a steady supply of clicks nobody meant to make, concentrated in exactly the placements that look cheapest. The signature is easy to spot and easy to miss if you only read the top line: a high click rate, a very short time on the landing page, and almost no conversions.

So set the primary metric past the click before the campaign starts.

• A post click engagement event: a scroll depth, a video start, time on page past a threshold, anything that requires intent.
• Cost per engaged visit, not cost per click, which makes expensive quality inventory and cheap accidental inventory comparable for the first time.
• Placement level reporting, with the right to exclude. Mobile app inventory is long tailed, and a small number of placements usually account for most of the waste. If a seller cannot give you placement level data and a block list, that is the thing to negotiate rather than the price.
• A view through comparison, because in feed native is often doing an awareness job while being measured as a direct response one.

Viewability is a floor, not a result

The industry's viewable impression standard asks for half of the advertisement's pixels to be on screen for at least one continuous second, and two seconds for video. On a desktop page that is a reasonable proxy for an advertisement having had a chance. In a fast moving mobile feed, a unit can clear that bar on the way past and never be read by anybody. Treat viewability as a billing hygiene check, which is what it is, and keep the judgement of whether the advertisement worked on reach, frequency and the engagement measure above.

The landing page is half of what you bought

A native click from a feed is the least committed click in advertising. The reader was doing something else, they are on a mobile connection, and they will give the page about as long as it took to tap it.

• Match the page to the unit. The headline they tapped should be the first thing they read. A click from a specific claim landing on a generic homepage is the most common way this money is wasted.
• Keep it fast and light on a real connection, tested on a mid range phone rather than on your own.
• No pop up over the content they just asked for. An interstitial on arrival undoes the one advantage native has, which is that the reader chose this.
• Ask for one thing. One field, one button, one decision. A form designed for a desktop visitor who came through search is the wrong form for this traffic.

Labelling, in one paragraph

Native formats exist because they match their surroundings, which is exactly why the label is not negotiable. The United States Federal Trade Commission's native advertising guidance and its endorsement guides govern it, the UK has its own rules under the advertising codes, and the platforms add their own labels on top. The practical requirement is the same everywhere: the label is above or at the start of the unit, in words an ordinary reader understands, and still visible when the unit is rendered small. The detail on labelling and the formats it applies to is in native content examples.

Native advertising on mobile: what buyers ask

Is social in feed advertising native advertising?

Yes, by the IAB's own definition: it is an in feed unit styled to match the surrounding content and labelled as paid. It gets discussed separately because each platform sells it in its own auction with its own creative tools, but the format and the craft are the same.

What does mobile native cost?

It is bought on cost per thousand impressions or cost per click depending on the route, and the range is wide: programmatic app inventory is the cheapest per thousand, premium publisher placements and branded content the most expensive. Compare routes on cost per engaged visit instead, which is the only figure that makes them comparable.

How is it different from a banner?

A banner is a finished image you supply, placed in a slot around the content. A native unit is components you supply, rendered into the content flow by the publisher's template. That is why a banner can be designed and a native unit has to be specified.

Which platform should a small advertiser start with?

The one where you can already see your audience and where the creative tooling is free, which for most businesses means a social platform rather than programmatic app inventory. Start where the reporting is good enough to tell you whether the clicks were real.

How much creative variation do I need?

More than for a banner, and of a different kind: several headlines and several images rather than several designs, because the platform is combining them. Three headlines and three images is a sensible opening set, with the losing combinations switched off rather than edited.

Does native advertising help with business to business buying?

It works where the audience is reachable by job rather than by behaviour, which on mobile mostly means professional social feeds. The formats and the two ways of running them are covered in sponsored content on LinkedIn.

The feed with no inventory in it

Native advertising is built on a simple insight: people accept a message that arrives in the shape of the thing they were already doing. The newest version of that shape has no advertising slot at all. Somebody asks an assistant on their phone which product or supplier to use and gets a short answer naming two or three, in exactly the register of the rest of the conversation. There is nothing to buy, no template to supply components to, and no placement report. What decides the answer is what can be retrieved about each company and whether it is consistent. Any advertiser spending on mobile reach is paying to be remembered at that moment, which makes it worth knowing what the answer currently says. AI Knows Us asks those questions across the main assistants and reports which companies get named and what the answers were built from.