Website Ad Formats Compared: Earnings vs Experience

Every ad on a web page is a trade. The unit earns something, and it costs the reader a little attention, a little load time, or a little patience. Some formats are cheap on that second side. Others look generous on a revenue report and quietly chew through your traffic.

This guide goes through the five formats most site owners choose between: display banners, in-article native units, sticky or anchor ads, interstitials and outstream video. For each one we look at how it earns, what it costs the reader and your Core Web Vitals, and where it fits. Dollar figures are estimates from the same benchmark model as the calculator, not quotes from any network.

Where a page's earnings come from

The calculator on this site starts from a reference page: four ads, 60% viewability, and the three everyday formats switched on (you can toggle them in the format cards). It then splits that reference RPM between the formats, and treats video and interstitials as extra layers on top. The table shows the weights it uses. If RPM as a term needs a refresher, see eCPM in plain English, and the methodology explains how the benchmarks were built.

To see what that means in practice, take a Tier 1 reference RPM of $17.00, which is the benchmark figure for Mediavine. Display contributes $10.20, native $3.74 and sticky $3.06, which adds up to $17.00. Switching on video adds a modelled $6.80 and interstitials another $2.55, for $26.35 in total. Real pages rarely reach that sum, because the extra formats compete with the others for the same advertiser budgets. Treat the additions as a ceiling to compare with, not a forecast. Video and interstitials are also not available on every network (check your own program's current terms).

Display banners: the workhorse

Display units are the rectangles and leaderboards you already know. They earn through volume: a page can carry several, and almost every advertiser buys them. That is why they carry the largest share of the model, and why they are the format you will adjust most often.

Their weakness is that they are the easiest to ignore. Readers have trained themselves to skip sidebars. A banner earns best when it is visible for a while, in a spot where eyes already rest, and it earns very little in a footer. Keep the count moderate and the sizes standard. Odd sizes get fewer bids.

In-article native: earned by being read

An in-article unit appears between paragraphs and takes on the typography of the page. People look at it because they are already reading that spot. That tends to give it better viewability than a sidebar unit, and the price per view often reflects that.

The cost is rhythm. Place one every few paragraphs and the article starts to feel like a brochure. Place the first after the opening section instead of the first paragraph, space them out, and keep them away from headings and images. A reader who has settled into the text forgives one unit far more easily than one placed in the first line.

Sticky and anchor units: easy views, easy overreach

A sticky unit stays pinned to the bottom or side while the page scrolls. Because it is on screen the whole time, its viewability is high, which advertisers like. It is also the format most likely to annoy a reader on a phone, where it can take a noticeable strip of the screen.

Use one sticky unit at most, keep it slim, and give the reader a way to close it if the network allows. Check that it reserves its own space so it does not push content as it loads, since that is where layout shift comes from.

Interstitials: a large payment with a large risk

A full-page interstitial between navigation steps can pay well per impression, which is why the model credits it with an extra 15%. It also sits right in the reader's way. Google's search guidance has long discouraged intrusive interstitials that cover content right after a visitor arrives from search, so one shown on the first page of a session is the riskiest placement you can choose.

If you use one at all, cap it at once per session, show it only after a second pageview, and never on the landing page. A better test is to ask whether the same reader would come back after seeing it. If you cannot answer yes with data, leave it off.

Outstream video: the extra layer with the biggest bill

Outstream players sit inside content and play when they scroll into view. Video demand is deep, and the model gives it the biggest add-on, at 40% of the reference RPM. The bill comes in three parts: heavy scripts that slow the page, audio or motion that annoys people, and layout changes if the player collapses when no ad is available.

Many small sites cannot run it at all, since outstream is mostly offered through managed or optimizing networks. If yours qualifies, start with a single player in a long article and check page speed before and after. Video earns well on pages people read to the end and poorly on short ones.

What the ads do to Core Web Vitals

Google measures three things that ads can damage. Largest Contentful Paint is how quickly the main content appears, and a good result is 2.5 seconds or less. Cumulative Layout Shift is how much the page jumps around, and a good result is 0.1 or less. Interaction to Next Paint is how quickly the page responds to taps, and a good result is 200 milliseconds or less. The overview of the vitals explains how field data is collected, and Google's Search documentation describes how they relate to ranking.

Layout shift is the usual ad problem. A slot that is empty until the ad arrives pushes the text down. Fix it by giving every slot a fixed minimum height, so the space is held whether or not an ad fills. LCP suffers when ad scripts compete with your hero image, so lazy-load units below the first screen. INP suffers from heavy video and many auctions on the main thread. Measure before you add formats, and again after each one.

A starting mix by site stage

There is no universal layout, but there is a sensible order in which to add things. The table gives a starting point for each stage.

Add one format at a time and compare revenue per session, not just RPM, because a unit that raises page RPM can still lower pages per session. The AdSense or Mediavine guide covers when the move to a managed network makes sense, RPM by niche gives context for the numbers above, and seven ways to earn more covers the wider set of levers. To test your own page count and viewability, use the AdSense calculator, the Mediavine calculator, or the main calculator.