Microsoft Ads Performance Max Previews: A QA Workflow

Your image, headline, logo, and call to action can each look fine in isolation and still produce an awkward ad when Microsoft’s automation puts them together. Until you inspect those combinations, creative approval is only half finished.

Microsoft Ads now gives Performance Max advertisers a practical way to close that gap. Ad Preview Hub shows eligible formats across devices and placements from within an asset group, then lets you share the preview with reviewers. Used properly, it becomes a creative quality-control step rather than another screen someone glances at before launch.

Decide what the preview can actually approve

A Performance Max preview answers a narrower question than many approval teams assume: can the assets in this group form acceptable ads across the formats currently available for review?

That is different from asking whether the campaign will perform. A polished preview cannot tell you which combination will earn the strongest response, where Microsoft will deliver most impressions, or whether the offer will convert. Those questions require live campaign data.

The preview can help you make concrete creative decisions before spending begins:

• Does every visible combination communicate one coherent offer?
• Can the headline, image, logo, and call to action be understood when Microsoft rearranges their roles?
• Does the creative remain recognisable across the devices and placements shown?
• Are claims, qualifiers, branding, and offer details consistent?
• Would a reasonable reviewer know what the advertiser wants the audience to do next?

Keep two approvals separate. The pre-launch approval confirms that the creative system is safe and coherent. The post-launch evaluation determines whether that system performs. If you combine those decisions, attractive mockups can acquire more authority than they deserve.

The word eligible also matters. Review everything the Hub makes available, but do not treat a set of previews as a promise that you have seen every possible impression. Your standard should be robust assets, not merely acceptable examples.

Run the review at the asset-group level

Open the relevant Performance Max asset group and select Preview ads. Work through the eligible formats, devices, and placements shown. If the campaign contains several asset groups, repeat the process for each one; an approval for one group says nothing about the combinations in another.

Before opening the preview, write a one-sentence brief for the asset group: who it addresses, what it offers, and what action it asks for. That sentence gives every reviewer the same standard. Without it, feedback tends to collapse into personal preferences about colour, wording, or imagery.

Then review in four passes:

1. Intent pass: Confirm that each preview still matches the asset group’s audience, offer, and desired action. If one combination appears to advertise a different product, promotion, or stage of the journey, the group is carrying conflicting jobs.
2. Combination pass: Read every displayed combination literally. Look for headlines that depend on a particular image, descriptions with unclear words such as “this” or “it,” repeated phrases, conflicting promises, and calls to action that do not fit the surrounding message.
3. Placement pass: Check the previews across every device and placement the Hub exposes. Inspect whether the brand remains identifiable, the focal subject remains understandable, the message hierarchy survives the layout, and important meaning depends on text embedded inside an image.
4. Risk pass: Verify names, offer terms, claims, qualifiers, required disclosures, brand treatment, and destination intent. Legal or policy-sensitive language should be reviewed in the assembled ad, not approved solely in the original copy document.

Do not stop after finding one polished render. Performance Max assembles assets across multiple placements, so the useful test is whether the group remains coherent when the presentation changes. One hero preview is evidence that one arrangement works. It is not approval of the asset system.

Fix the reusable asset, not the individual screenshot

When a preview looks wrong, it is tempting to describe the visible symptom: the logo feels small, the copy seems repetitive, or the image does not make sense beside that headline. The more useful question is which asset created the dependency.

Use four diagnostic questions:

• Can the text stand alone? A headline that only makes sense beside one particular image is fragile in an automatically assembled campaign.
• Can the image support more than one line of copy? If its meaning changes completely when paired with another eligible message, it may be too narrowly constructed for the group.
• Do all calls to action point toward the same next step? An asset group should not make the audience alternate between incompatible actions.
• Do the assets belong to the same audience and offer? If the preview exposes several propositions competing for attention, the problem may be asset-group scope rather than visual execution.

Rewrite or replace the asset that fails those tests, then preview the group again. Do not approve a weak asset because it happened to receive a helpful companion in one render. Microsoft’s assembly process may place it in a less forgiving context.

Separating assets into another group can be appropriate when they genuinely represent a different audience, offer, or creative concept. It should not be used merely to hide an asset that cannot communicate clearly. The cleaner rule is simple: every asset in a group should contribute to the same decision, even when its neighbouring assets change.

After a material edit, reopen Preview ads and repeat the affected passes. Approval belongs to the current collection of assets and the ads they can form, not to an old screenshot or copy deck.

Turn the shareable link into a controlled approval

Ad Preview Hub can generate a secure link for stakeholders or clients. Reviewers do not need access to the Microsoft Advertising account, and Microsoft says the preview is accessible only to the person who created the link and the people with whom it is shared.

That removes account-access friction, but it does not create an approval process by itself. Sending a bare link invites vague comments and makes it difficult to determine what was actually approved.

Send the link with five pieces of context:

• Scope: Name the campaign and asset group under review.
• Version: Identify the creative round or date so reviewers do not approve an obsolete set.
• Intent: Include the one-sentence audience, offer, and action brief.
• Reviewer lens: Tell each person what they own, such as brand treatment, claims, commercial accuracy, or channel execution.
• Decision: Ask for one of three outcomes: approved, approved after named corrections, or blocked with a specific reason. Include an owner and deadline.

Request feedback in a form that can be acted on. “Make it pop” does not identify a failure. “The product name is unreadable in the mobile preview” identifies the context, symptom, and asset likely to need attention.

Give one person responsibility for consolidating comments. Otherwise, a copy change requested by one reviewer can invalidate a brand or compliance approval already given by another. Once revisions are complete, circulate the current preview and ask reviewers to confirm the final state.

Keep the decision in your normal project record, even though the preview link is secure. Record the asset group, version, approvers, unresolved exceptions, and approval date. A review surface helps people see the ad; it does not automatically replace the audit trail your team may need later.

Key takeaways for your Performance Max launch gate

• Open each asset group and select Preview ads; do not approve an entire campaign from one group’s previews.
• Inspect every eligible device, placement, and format shown rather than choosing the most attractive example.
• Reject assets that only make sense beside one specific companion asset.
• Check audience, offer, action, claims, and brand consistency in the assembled ads.
• Send the secure preview link with scope, version, reviewer responsibility, decision options, and a deadline.
• Record approval outside the preview and rerun the review after material creative changes.
• Use previews to validate creative coherence, not to predict campaign performance.

Put this workflow on the next asset group scheduled for launch. If a reviewer can identify the offer, audience, and next action across the available previews—and no asset depends on a lucky pairing—you have a defensible creative approval. Let the live campaign data answer the separate question of what performs.

References

• Search Engine Land — Microsoft Ads adds Ad Preview Hub to Performance Max