All guides · Ad formats · October 13, 2026

Compare product variants fairly in AI ads

A comparison ad helps a viewer choose between real options. Decide which supplied difference matters: colour, visible shape or a verified specification. Keep viewing conditions similar so the ad does not make one option look preferable simply by changing the light, camera distance or product size.

Choose a comparison the evidence supports

Hypothetical example: a shop supplies the same notebook design in blue and green, photographed from matching angles. The two colour choices are confirmed. Page count, material and size differences are unknown. The comparison should be about those colours, without calling one more durable or larger.

If the products have different dimensions, obtain the actual measurements or a real photograph showing their relationship. Two separately generated close-ups do not establish relative size. An apparent larger product might simply be nearer to its camera.

Make separate references for separate variants

Do not ask the model to recolour a single reference from memory. Each variant needs its own supplied images and identifying facts. The colour-variant guide explains how to keep shades separate.

fairytale video selects one product for an ad. Prepare a separate ad for each variant with matched scene, camera and action direction; preserve the finished first version when making the other. Compare the completed versions before choosing a presentation for publication. If you assemble a side-by-side presentation elsewhere, label each option clearly and retain the original files.

Write a matching-shot specification

Hypothetical comparison setup:
Same photographed desk and neutral daylight.
Same notebook position, orientation and camera framing.
Same small camera move and stationary product.
Blue reference in one ad; green reference in the other.
No changes to exposure, props or claimed specifications.
Comparison message: two supplied cover colours.

Matching direction does not guarantee identical outputs. Inspect the finished shots for changes that would undermine the comparison, particularly perspective and colour. Do not adjust one shade until it is more appealing but less faithful to its source.

Separate choice from performance

• Colour choices can be illustrated from accurate references.
• Mechanisms can be shown when their states are supplied.
• Capacity claims need confirmed dimensions or real fit evidence.
• Durability, heat, protection and outcome comparisons need real proof.

A generated demonstration should not become a “winner” test against a competing product. Keep the copy to verifiable distinctions rather than inferred superiority.

Review and preserve the comparison

Check both versions against their original photos and ask whether a viewer can identify the intended difference without hearing the line. Use scale review for size questions and creative testing when comparing how different ads perform. Product choice and ad performance are separate questions, and each needs its own evidence.