Native advertising

Native advertising, in the affiliate sense, is paid placement in content recommendation widgets: the "you may also like", "around the web" and "recommended for you" blocks under and beside articles on news, entertainment and lifestyle sites. Each ad is a thumbnail and a headline that looks like another article. The click goes to a page you choose, usually an advertorial.

The attention comes from people reading content. They are not searching and not scrolling a social feed; they have finished or paused an article and are open to reading another. That mindset is why native suits long-form persuasion: an advertorial story reads naturally after a news article, and older, less ad-savvy audiences click more readily.

The inventory belongs to publishers, who install a native network's widget and share the revenue. Networks aggregate thousands of sites and sell the clicks to advertisers. For affiliates, the relevant networks differ mostly in how strict they are. MGID is widely used for nutra and other affiliate verticals. Taboola and Outbrain have larger premium publisher networks and considerably stricter content rules, which make them hard to use for most nutra offers.

Native is a middle ground among paid sources: more expensive per click than push, much less restrictive than Facebook, and with traffic quality that varies enormously from one publisher site to the next.

Native is bought almost always on a cost per click basis, in an auction.

You create a campaign with targeting (country, device, OS, sometimes browser and language), upload several ads (image plus headline), and set a CPC bid. The network decides where and how often your ads show by combining your bid with the ad's expected CTR: a better-performing ad wins more impressions at the same bid, because the network earns only when someone clicks. You pay for clicks, not impressions, so a low-CTR ad simply gets little delivery.

The part that makes native distinct is placement-level bidding. Your campaign runs across hundreds or thousands of publisher sites and widgets, and performance differs sharply between them. Networks expose the placement (a site or widget ID) in reports and as a URL macro, and let you blacklist placements, whitelist them, or adjust the bid per placement with coefficients. Most of the work in native is this pruning: start broad, cut placements that spend without converting, raise bids on the ones that pay.

Budgets are set per campaign, daily and total. Networks usually require a prepaid deposit. MGID states a minimum account deposit of $100, a minimum campaign budget of $50 and a recommended starting deposit of "$650+" (MGID blog, updated August 2026).

Every ad and landing page goes through moderation before running. Expect it to take longer and reject more often for health offers than for other verticals. MGID's help centre says moderation takes "up to 24 hours for teasers in English and up to 72 hours for advertisements in other languages", can run longer when documents or a security check are needed, and continues on weekends and most public holidays; after 72 hours it suggests contacting your account manager (MGID help centre, as of September 2026).

The core format is the widget ad: an image (usually landscape or square), a headline of limited length and sometimes a short description or brand name. The ad sits in a grid of similar-looking items, so the thumbnail and the headline carry the click. Some networks also sell in-article placements, video and interstitial formats, but affiliates mostly buy widgets.

Targeting options typically include:

• Geo: country, often region or city.
• Device and OS: desktop, mobile, tablet; iOS, Android, Windows.
• Browser and language.
• Publisher category: news, entertainment, health and so on.
• Placement: blacklists and whitelists of sites or widgets, and per-placement bid adjustments.
• Time of day and day of week.
• Audience or interest segments and retargeting, where the network offers them.

For nutra, the most useful levers are geo, device and placement. Health offers usually convert best on mobile in countries with COD, and older audiences on news sites often outperform younger ones on entertainment sites, but only testing tells you that for a given offer.

Creative rotation matters: a native ad's CTR decays as the same readers see it repeatedly, so plan for new thumbnails and headlines every few days to weeks. Note the launch date of each creative so you can see when its CTR starts to fall.

Native networks moderate both the ad and the landing page, and health products are where most rejections happen. Rules differ by network, so read the one you buy from; the patterns below come from their published policies.

Taboola (its advertiser platform is now branded Realize) publishes a health and beauty FAQ (last updated May 28, 2025) that lists what gets rejected (Realize help centre):

• "Unrealistic or photoshopped before and after images."
• "Falsely claiming that the product is recommended by a healthcare professional."
• Weight loss claims such as losing "2 lbs or more a week for a month or more without dieting or exercise".
• Absolute cure or prevention claims, and content discouraging conventional medicine.
• Claims without support: it asks for linked scientific studies behind health claims.

Its landing page rules require "Advertorial" written clearly at the top of advertorials, forbid giving the impression of a news site, require paid testimonials to be labelled, disallow celebrity names and images without permission, and warn against countdown timers that create false urgency (Realize landing page policies).

MGID is more permissive in practice, but its compliance checklist tells advertisers to "avoid endorsing food supplements by doctors, pharmacists, or dietitians with no legal authorization", not to claim "clinically tested" without proof, to label advertorials, and not to use fake buttons, back-button redirects or "are you leaving?" popups (MGID blog, 2020). Its advertiser terms forbid cloaking (MGID Advertisers TOS).

The checklist dates from 2020, but MGID's current help centre says the same things in more general terms. It lists health and wellness products as restricted ("must avoid misleading claims, exaggerated results, or unsafe practices"), supplements and herbal remedies as restricted ("must comply with local regulations and avoid medical claims"), and "medical or health-related claims that are misleading, unsafe, or unverified" as prohibited (MGID, restricted and prohibited content, as of September 2026). A separate page bans "guaranteed results", fabricated testimonials and endorsements, imitation of media outlets and public figures, false countdown timers and scarcity messages, and AI-generated "false representations of people, products, or results" (MGID, prohibited practices, as of September 2026).

In short: fake doctors and celebrities, miracle cures, doctored before and after photos, unlabelled advertorials and fake urgency are what gets rejected.

MGID's own cost guide gives these industry-wide native benchmarks: CPC "$0.20 to $0.50", CPM "$5 to $15" and CTR of 0.30 to 0.60%, and it names nutra among the verticals with "strong advertiser demand" (MGID blog, updated August 2026). These are a vendor's averages across geos and verticals. Tier 3 countries are far cheaper per click, tier 1 countries far more expensive, and nutra competes with other high-paying verticals for the same placements.

The same guide suggests testing on "5,000 to 10,000 clicks". At a $0.05 CPC in a cheap geo that is $250 to $500; at $0.30 in a tier 1 geo it is $1,500 to $3,000. Plan the test budget from the CPC you actually see in the network's traffic estimator for your geo, not from averages.

For a nutra COD offer paying around $25 to $30 per approved lead, the arithmetic usually requires cheap clicks and a strong prelander: at $0.05 per click you need an EPC above $0.05, which means roughly one approved lead per 500 to 600 clicks.

Before testing, set a budget per geo that you can afford to lose, and decide how many clicks a placement gets before you block it.

Native tracking revolves around the placement ID, because placements are what you optimise.

In your tracker, add the native network from its template. The destination URL of each ad then carries the network's macros: the placement or widget ID, the ad or teaser ID, the campaign ID, the network's click ID, and where supported the cost per click. Check the raw click log after launch to confirm each one fills in; a placement column full of the literal macro text means the template or the macro name is wrong.

From the tracker to the affiliate network, pass the tracker's click ID in a sub-parameter and set up the approval-aware postback as described in tracking and attribution.

From the tracker back to the native network, set up a conversion postback using the network's click ID, so the network can optimise toward converting placements. Most trackers ship a ready template for MGID with these macros filled in, and MGID's help centre has setup guides for Keitaro, Binom, Voluum and RedTrack among others (MGID help centre). Importing costs through MGID's API is a different matter. Voluum documents a full MGID integration that pulls costs and campaign statuses from MGID's API and overrides the cost set in the campaign (Voluum docs), and RedTrack lists MGID among its API integrated traffic channels (RedTrack help). Keitaro's documentation lists API cost imports for Facebook but not for MGID, and Binom describes its traffic source integrations as a way to start and pause campaigns; with either, MGID costs arrive per click through the cost token in the campaign URL or through a manual cost update (Keitaro docs, Binom docs). All as of September 2026.

Finally, match time zones between the network and the tracker, and reconcile clicks weekly. A persistent gap between network clicks and tracker clicks on one placement is a signal to look at that site's traffic quality.

Native fits offers that need to be explained before they are bought, and audiences that read.

• Nutra, especially joints, cardio, vision, hearing and weight loss, where the advertorial story does the selling and the older news-reading audience matches the buyer. See nutra.
• Finance lead generation, insurance and home services in tier 1 countries, on the stricter networks.
• E-commerce and "gadget" offers sold through review-style prelanders.
• Content arbitrage: buying native clicks to your own articles and earning from display ads, a separate business model with thin margins.

Native is a poor fit for impulse offers that need no explanation (push or pop are cheaper), for offers whose only angle breaks health or finance rules, and for very small budgets, since placement-level optimisation needs thousands of clicks before patterns appear.

Where native budgets burn:

• Running broad for too long. Without regular blacklisting, low-quality placements absorb most of the budget. Review placements daily during a test.
• Bot and made-for-advertising sites. Some publisher sites exist to resell traffic and produce clicks with no buyers. High CTR with zero leads is the signature.
• Winning placements that disappear. Top widgets change owners, layouts or policies. A campaign resting on three placements is fragile.
• Creative fatigue. A thumbnail that worked last week slowly loses CTR, and the auction punishes low CTR with less delivery.
• Advertorials that break the network's rules. Rejections cost days; a campaign stopped after launch for policy reasons can also lose the placements it had earned.
• Buying premium networks with grey angles. Taboola-level rules and grey nutra prelanders do not mix, and the result is rejected ads, not cheaper clicks.