Optimal Format Selection in AI Max for Shopping: an implementation playbook

Short answer: Implement Optimal Format Selection only after the Shopping campaign has reliable feed data, destination pages and clear measurement boundaries. Google says the feature can automatically choose between text and Shopping ad formats based on what is most relevant to a shopper's needs. That is a delivery capability, not a guarantee that automation will outperform every manually segmented setup.

What the feature does

Google's 2026 AI Max for Shopping announcement says Optimal Format Selection can automatically choose the format—text or Shopping ads—based on what is most relevant to the shopper's needs.

That creates a mixed-format delivery layer inside a Shopping campaign.

The implementation challenge is not simply turning it on. It is ensuring that both possible formats have accurate inputs, compliant creative boundaries and destinations that support the intent being matched.

Step 1: verify campaign eligibility and scope

Before rollout, document:

• account;
• campaign;
• country/market;
• language;
• product scope;
• bidding strategy;
• conversion goals;
• existing product targeting controls;
• whether Final URL Expansion is enabled;
• current text customization state.

This becomes the pre-change baseline.

Step 2: prepare Shopping-format inputs

The Shopping side depends on reliable Merchant Center data.

Verify:

• product identity;
• title;
• image;
• price;
• availability;
• variant logic;
• descriptive attributes;
• product URL;
• policy status.

A format-selection system cannot repair factual product-feed errors.

Step 3: prepare text-format inputs

If the system can choose a text experience, the campaign also needs strong textual constraints.

Review:

• brand terms;
• prohibited claims;
• required disclaimers;
• price language;
• offer language;
• product-category wording;
• market/language constraints;
• destination compatibility.

Do not assume text generation will infer the advertiser's legal or brand boundaries correctly without explicit controls.

Step 4: verify destination logic

Optimal format and Final URL Expansion can interact with destination selection.

Check:

• landing-page product relevance;
• price and stock consistency;
• canonical state;
• mobile quality;
• product/category-page suitability;
• legal/support/careers exclusions where relevant;
• tracking parameters;
• conversion tags.

A text ad landing on the wrong page can create a worse experience than a less-flexible format.

Step 5: define format-observation fields

Reporting should let the team understand how delivery changed.

Where available, store:

• campaign;
• date;
• query/search-term theme;
• product/category;
• observed format;
• destination;
• impression;
• click;
• conversion;
• conversion value;
• cost;
• policy or brand issue.

If the platform does not expose a field directly, do not infer it and present the inference as observed fact.

Step 6: build comparable cohorts

Do not judge format selection from one high-performing query.

Use cohorts such as:

• same product category;
• similar price band;
• same market;
• stable inventory;
• comparable promotion state;
• similar campaign learning stage.

Compare trends across enough observations to avoid one-off conclusions.

Step 7: separate relevance from incrementality

The system may choose a format that appears more relevant to the shopper. That does not automatically mean the format caused incremental sales.

Keep layers separate:

Delivery

Which format and destination served?

Engagement

What click or interaction occurred?

Conversion

What attributed conversion followed?

Incrementality

Did the automated format choice cause a result that would not otherwise have occurred?

The last question requires stronger experimental evidence.

Step 8: monitor format-specific failure modes

Potential issues include:

• text ad uses unsupported product claim;
• Shopping ad shows stale price;
• text format lands on a broad category page when a product page is needed;
• product image/variant does not match user intent;
• format changes complicate reporting;
• tracking differs across destinations;
• brand or legal wording is inconsistent.

Assign an owner to each failure class.

Step 9: define rollback conditions

Pause or narrow automation when:

• destination quality drops materially;
• price or availability mismatch appears;
• generated text violates approved constraints;
• tracking breaks;
• policy issues increase;
• measurement cannot distinguish the new delivery state;
• business-critical products require tighter control.

Keep the pre-change campaign state documented.

Step 10: scale only after operational validation

A conservative rollout can be:

1. one bounded campaign cohort;
2. stable product feed;
3. verified destinations;
4. approved text controls;
5. observation period;
6. compare delivery and business outcomes;
7. expand only if failures remain controlled.

Do not scale merely because activation is available in one click.

The implementation rule

Optimal Format Selection should be treated as automated delivery choice inside a governed campaign system.

Prepare both Shopping and text inputs, verify destinations, preserve format-specific evidence and use experiments for causal claims. The feature can choose a format; the advertiser still owns truth, compliance and interpretation.

Sources reviewed

• https://blog.google/products/ads-commerce/ai-max-for-shopping/