Merchants keep asking for a single canonical list of ad dimensions, but the more useful question is whether the page hosting those ads is even accessible, since the errors that break screen readers and the errors that break banner rendering come from the same sloppy markup.
Most merchants type “ad sizes reference” expecting a chart of pixel dimensions for Meta and Google Display units, and then build a campaign around whatever list loads first. The gap in that approach is not the dimensions themselves, it is that nobody checks whether the page receiving the traffic can actually carry an ad well. A banner sized correctly on a page that fails basic accessibility checks still loses a share of visitors before the creative even registers, and that loss has nothing to do with pixel width.
The measured facts here come from WebAIM’s Million project, an accessibility scan of one million home pages, and the current run puts WCAG conformance at 95.9%, up from a prior reading of 94.8%. That sounds like broad improvement, but the same measurement period recorded a 10.1% increase in the number of detected errors. Both numbers are real and both come from the same source, so the honest reading is that more pages are being scanned and more defects are surfacing even as the percentage that pass basic conformance improves. A banner reference chart that ignores this context is describing pixels on a canvas that, for a meaningful share of sites, is already failing its visitors before an ad loads.
Three specific defects from the WebAIM data map directly onto ad performance. Low contrast text appears on 83.9% of measured pages, and any ad copy or call to action rendered on top of a low contrast background inherits that problem. Missing alternative text for images affects 53.1% of pages, which matters directly for image-based ad units since a banner without alt text is invisible to a screen reader user. Missing form input labels show up on 51% of pages, relevant whenever an ad’s destination is a signup or checkout form rather than a product page.
| Measured defect | Share of pages affected |
|---|---|
| Low contrast text | 83.9% |
| Missing alternative text for images | 53.1% |
| Missing form input labels | 51% |
| WCAG conformance overall | 95.9% |
No confirmed pixel dimensions for Meta or Google Display units appear in this article’s fact set, so none are printed here. A reference site’s only job is to report what has been checked, and printing a remembered size chart without a dated source is exactly the kind of error that compounds across the web the same way a missing alt text tag does. The numbers in this piece rest on the WebAIM Million scan as loaded on September 9, 2026, covering one million home pages, and on USPS Notice 123 as loaded the same day for the unrelated postal rate figures included in the source set. Neither document specifies digital ad dimensions, and neither should be stretched to imply one.
USPS Notice 123 sets a 5% rate adjustment for authorized publications and a 1% nonadvertising adjustment factor for mailed periodicals. These figures describe physical mail pricing and are unrelated to digital ad sizing, but they are part of the same verified fact set behind this article, so they are named here rather than omitted silently. A reference site that hides what it does not know is less useful than one that says plainly where the boundary of its verified data sits.
Check the platform’s own specification page before locking any campaign to a dimension, and treat the page hosting the ad as part of the campaign, since a low contrast or unlabeled page undermines a correctly sized banner just as much as a wrong pixel count would.
This reference does not carry a verified list of pixel dimensions in its current fact set, because no dimension numbers were confirmed against a primary source at the time of writing. Rather than print a size chart pulled from memory, the article sticks to what is actually measured, which is the accessibility state of the pages that host these banners. A follow-up piece can add dimensions once they are checked against Meta's and Google's own specification pages. Until then, treat any size chart you see elsewhere as unverified unless it links to the platform's own docs.
Because 53.1% of measured homepages are missing alternative text for images, and ad banners are images or image-like elements dropped into that same markup. Missing form input labels affect 51% of pages, which matters when an ad links to a signup or checkout form. Low contrast text sits at 83.9%, and ad units inherit the visual weaknesses of the page they sit on. A banner rendered correctly on a page that fails these checks still delivers a poor experience to a share of visitors.
Both, depending on which number you look at. WCAG conformance moved up from 94.8% to 95.9%, which reads as improvement. At the same time the number of detected errors increased 10.1%, meaning more pages, more content, and more errors are being found even as the conformance percentage rises. The two figures come from the same measurement round, so neither one tells the whole story alone.
No, and it should not be confused as such. Notice 123 covers physical mail pricing, including a 5% rate for authorized publications and a 1% nonadvertising adjustment factor. It is listed here only because it is part of the same measured fact set this reference draws from, not because it defines any digital banner dimension.
Go to the platform's own specification page rather than a secondhand list, since dimensions change and a stale chart causes rejected creative. This article's fact set does not include confirmed pixel values for Meta or Google units, so none are printed above. Once such values are checked and dated, they will be added with a source link and a checked date, following the same rule every other number on this site follows.