Microsoft Ads Adds Ad Preview Hub to Performance Max, With 30-Day Public Share Links

The preview button shipped. So did a public link.

Microsoft Advertising's Ad preview hub has reached Performance Max. As of July 28, 2026, Microsoft's own help article documents a Performance Max preview step during campaign creation, a publisher dropdown that now includes Bing Search alongside MSN and Outlook, and a Share button that produces what the article calls "a public, view-only link that expires after 30 days." Two details in that sentence deserve more attention than they have been getting. The rollout is partial. And public means public.

What Microsoft's documentation actually says

The announcement travelled through a liaison post and a help-doc update rather than a monthly product blog, which is how the caveats got lost in transit. The documentation is more careful than the summary of it.

The rollout is staged, in Microsoft's own words

The article carries a note before the first paragraph: "This feature is currently available for Audience campaigns. Some advertisers are enabled for Performance Max campaigns, but not everybody has this feature yet." The Performance Max section repeats the point on its own line: "Not everyone has this feature yet."

That is two separate warnings in one document, and neither reads like general availability. The practical consequences are small but real:

• If you open a Performance Max asset group this week and find no preview control, nothing is misconfigured and there is no setting to hunt for
• You are simply not in the enabled cohort yet, and Microsoft's own line is that it is coming
• Plan the workflow around it, but do not promise a client the artifact on a specific date

Where the preview sits in the Performance Max flow

Microsoft is explicit about the entry point: "You can see your ad previews during Performance Max campaign creation." It arrives late in the build, after you have worked through the entire asset group:

• Asset group name, text guidelines and the final URL
• Images and logos
• Short headlines, long headlines and descriptions
• Business name, call to action and extensions

Only then do you select Preview ads in the ad preview area.

From there the hub is a set of dropdowns rather than a single render:

• Device switches the render between Computer and Smartphone
• Publisher selects the surface your assets get composed into
• Asset steps through individual images when the group holds several
• Extension previews extension types alongside the base ad
• Full view drops the ad back into the context of the publisher's page instead of showing it in isolation

That last toggle is the one that earns the tool its keep. It is the difference between checking that an image looks good and checking that it survives being surrounded by someone else's content.

The hub is five months old, not five days old

The tool itself is not new. Microsoft introduced it in a product-news post dated February 17, 2026: "We've introduced an enhanced ad preview hub for Audience ads." That launch was narrow. The same post states that "The ad preview hub currently supports previews for MSN and Outlook. More sites will be added in the future," and closes with a commitment: "The new ad preview hub will expand to other campaign types in the near future."

July is that promise landing. Reading this as a brand-new tool inverts the story. The interesting part is not that Microsoft built a preview surface, it is which campaign type and which placement it reached next.

Bing Search is the part that is genuinely new

Strip away the framing and one capability actually changed. Search inventory became previewable, and only on the Performance Max path.

Two dropdowns, two different lists

The two documented flows expose different publisher options. In the Performance Max section, the Publisher dropdown offers MSN, Bing Search or Outlook. In the Audience campaign section, the same dropdown offers only MSN or Outlook.

Dimension	Audience campaigns	Performance Max
Publishers previewable	MSN, Outlook	MSN, Bing Search, Outlook
Devices previewable	Computer, Smartphone	Computer, Smartphone
Documented entry points	Creation and ad editing	Creation
Rollout status	Available	Partial
Share link	Public, 30 days	Public, 30 days

That is the whole delta, and it is a meaningful one. A Performance Max asset group can serve a search results page and a content feed from the same pile of assets, and until now the search half of that was invisible before spend.

The article contradicts itself, and that is worth knowing

The general placement line in the same document still reads: "We currently show ad previews for MSN and Outlook placements, with more publishers coming soon." That sentence sits above a Performance Max section whose dropdown includes Bing Search. The summary copy is Audience-era text that has not caught up with the flow beneath it.

Do not plan around the summary sentence. Plan around the flow you are actually standing in, and re-read the section that matches your campaign type. Documentation that expands in stages tends to be accurate in the steps and stale in the overview.

Why a search surface changes what you check

A content placement and a search results placement compose your assets into completely different containers. The same image that reads well as a large landscape unit inside an article page has to hold up in a compact slot beside text results, on a screen a fraction of the width. Three things behave differently there:

• Headline truncation, which bites much earlier in a narrow unit
• Logo legibility, which depends on the rendered size, not on your source file
• The crop of the focal subject, which the container decides and you do not

This is the same discipline that platform-specific formats already demand elsewhere. If you have never audited your creative against per-surface requirements, the video ad specs for every major platform is the reference that makes the exercise concrete rather than theoretical.

"Public, view-only" is not the same as private

Here is the line most worth reading twice, because it appears identically in both the Audience and the Performance Max share flows, and because it says the opposite of what a shareable approval link is usually assumed to be.

What the words actually mean

Microsoft's note beside the Share step: "This is a public, view-only link that expires after 30 days." Three properties, and only one of them restricts anything.

• Public means no Microsoft Advertising account is required, and no identity is checked at the door
• View-only restricts editing, not access
• Expires after 30 days bounds the exposure in time, not in audience

Put together, the protection model is an unlisted URL with a countdown on it. That is genuinely useful for its intended job, which Microsoft describes plainly: sending previews to "your team, stakeholders, or clients," including people who "do not use the Microsoft Advertising platform," so that "This streamlines the approval process to get your campaigns up and running quicker." It is not an access-control system, and it should not be treated as one.

The 30-day clock cuts both ways

The expiry cuts in both directions at once:

• In your favour: a link that dies on its own beats a permanent artifact nobody remembers to revoke, which is the failure mode of every ad-hoc method it replaces, from screenshots in email threads to files parked in a shared drive
• Against you: approval cycles at large brands routinely run longer than a month, so a link pasted into a planning doc in week one is dead by the time someone reopens that doc next quarter

If your approval trail needs to be auditable later, the link is evidence with a countdown on it. Archive the previews you will need to point at again.

A workable rule for client creative

Agencies should treat a preview link the way they treat an unlisted video, not the way they treat a permissioned document. That leads to three simple habits:

• Do not put unannounced products, dated promotional pricing or embargoed campaign work behind a link whose only defence is that the URL is hard to guess
• Assume any link you send can be forwarded outside the recipient list, because nothing in the mechanism prevents it
• Keep the link as a convenience layer on top of your real approval record, not as the record itself

If your approval process is currently held together by screenshots and message threads, this feature is a good prompt to formalize it. Building creative operations that scale covers the part Microsoft cannot ship for you.

What the preview does not cover

A pre-launch render check is valuable precisely because it is narrow. Knowing its edges keeps it from being oversold internally, which is how useful tools lose credibility.

Microsoft says the preview can drift from delivery

The article includes its own disclaimer: "Please note that ads may differ from their previews as placements update." Variations, it adds, may occur due to publisher format changes. Microsoft positions the hub as the tool that "lets you see what your ads will look like before your campaign goes live," and that is exactly the right claim. It is a snapshot of how your assets compose today, on the placements the hub currently renders, not a guarantee about a publisher layout that changes next month.

Creation-time, on the Performance Max path

There is an asymmetry between the two flows that is easy to miss. For Audience campaigns, Microsoft documents that "You can see your ad previews during Audience campaign creation or when editing your ads." The Performance Max section documents the creation path. As of July 28, 2026, the article does not describe an editing-time equivalent for Performance Max.

For a campaign type built around asset groups you revise continuously, that matters. Preview before you launch is a different workflow than preview whenever you swap an asset, and only one of those is documented today.

A render check is not a performance check

The most important limitation is conceptual. Two questions look similar and are not:

• What the preview answers: does this asset compose correctly, on this publisher, on this device, inside this container
• What it cannot answer: whether the angle earns attention, whether the hook holds past two seconds, or whether this asset beats the one you ran last month

A creative can render beautifully in all three publishers and still fail to earn a click, and a slightly awkward crop can sit on top of the best-performing angle you have.

Keep the two questions separate and staff them separately. How to run creative tests that actually tell you something handles the second question, and reading creative analytics without fooling yourself is what keeps the answers honest once results come in.

One master cut is the wrong unit of production

This is the second-order effect, and it is the reason a preview button belongs in a conversation about video production at all.

What breaks in exactly one placement

Per-placement preview is a detection tool, and what it detects is asymmetric failure. Most creative does not fail everywhere. It fails in exactly one container, and usually in one of three ways:

• The subject is framed for a landscape crop and gets clipped on a narrower unit
• The product sits precisely where the publisher's own interface chrome overlaps it
• On-screen text sized for a full-width article slot turns to mush beside search results

Before per-placement preview, that failure was discovered by performance data, weeks and a budget later, and it was almost always misdiagnosed as a bad angle rather than a bad fit. Now it is visible before launch, which changes what you are supposed to do about it.

The variant math, worked

Count the render contexts the documentation itself exposes:

• Three publishers on the Performance Max path: MSN, Bing Search and Outlook
• Two devices: Computer and Smartphone
• Six render contexts per asset, before you have added a single variant

Run two asset groups against the same catalogue and you are inspecting twelve. Add a second image variant per group and it is twenty-four.

The instinct at that scale is to compromise: find the one asset that is acceptable in all six contexts and ship it everywhere. That is the wrong move, because an asset optimized for the intersection of six containers is optimal in none of them. The right move is a variant per context that fails, and the only reason teams avoid it is that variants have historically been expensive. When each cut means another creator booking, another shoot day and another licensing window, what UGC creators actually charge explains why one master cut became the default in the first place.

Approval stops being the last step

There is a cultural shift hidden in the workflow change. A preview that arrives at the end of production is a rubber stamp. A preview that arrives before spend, with per-placement granularity, is a brief for the next round of assets.

Producing that next round has its own constraint, which is that six variants of one ad should still look like six variants of one ad. Holding a single presenter steady while everything around them changes is the workflow for keeping one AI actor consistent across variations, and it is what separates a per-placement set from six unrelated ads.

The same pressure is showing up across the ecosystem. Google's 2026 creative standards for video ads push multi-format coverage across 16:9, 1:1 and 9:16 as the default rather than the exception, and video campaign groups make coordinated frequency a native objective, which quietly multiplies the number of distinct creatives a plan needs. Automated, machine-assembled ad formats keep expanding the number of slots your assets have to fill, and every platform is now retrofitting the visibility layer that lets you see them before spend. Visibility without production capacity just means watching your creative break in higher resolution.

How Novoads solves the per-placement variant problem

Novoads is a global AI UGC video-ad generator: write or auto-generate a script, pick an AI actor, and get an ad-ready video in minutes, with voices available in 31 languages. For per-placement work, the part that matters is the output side:

• Video renders vertical, square or horizontal (9:16, 1:1 and 16:9)
• Product-to-ad images cover six aspect ratios (4:5, 9:16, 1:1, 16:9, 2:3 and 21:9) at 0.3 credits per generated image

That changes the arithmetic behind a variant. A UGC-style video runs roughly $2 to $11 depending on the model and length, so cutting a placement-specific version is a decision about ten minutes of work rather than a decision about a shoot day. When a preview shows an asset breaking in one container, you can generate the variant that fits it instead of negotiating a compromise everywhere else. Start for $1: the trial costs $1 for 3 days of access, then continues at $49 per month, and you can cancel anytime.

The preview is the cheapest place to change your mind

Every creative decision has a price that rises with how late you make it. Changing an asset in the preview screen costs a render. Changing it after launch costs the spend that taught you it was wrong, plus the reporting cycle that let you notice, plus the argument about whether the angle or the crop was to blame. Microsoft moving that decision point earlier for Performance Max, on a staged rollout with a public link and a documented drift warning, is a modest feature with an outsized implication: the industry is running out of excuses for shipping one master cut and hoping.

The advertisers who get value from this will not be the ones who preview more carefully. They will be the ones who treat every render context the preview exposes as a slot that deserves its own asset, and who can actually produce at that width. The button tells you what breaks. It cannot make you the creative to replace it.