How to create a 7-asset ecommerce ad campaign in Lamina

The quickest dependable way to build a seven-asset ecommerce campaign in Lamina is to approve one product-faithful hero, then use it as the continuity reference for three stills, one master vertical video, and two cutdowns. Don’t generate seven unrelated packshot ideas. Build one visual system, lock the SKU facts, and change only the scene or edit.

Keep the deliverable set tight: one hero product image, three lifestyle variations, a 10–15 second master vertical ad, a six-second hook cutdown, and an eight-second proof-and-CTA cutdown. Paid social gets a coherent launch package, without palette, packaging, copy, or camera treatment wandering between assets.

Use imagery you own or are authorized to market. The original product image is the SKU truth source; check every generated candidate against it before it reaches a PDP, Meta placement, or creator brief.

What needs to be ready before you upload a product image?

Build a SKU truth pack before you enter Lamina: the sharpest approved packshot, exact SKU name, color, material, dimensions, visible logo and label, cap or closure, product count, and approved factual claims. Weak references produce weak work. A clean front-facing image with readable packaging gives the reviewer a real comparison point.

Lamina’s workflow benchmark calls for locking the brand kit first, keeping a separate reference pack for every SKU or variant, putting a product-truth approval gate ahead of creative selection, and making channel variants only after approval. Treat a 250 ml amber bottle and a 500 ml refill pouch as separate references, even under one campaign name. Their proportions, labels, closures, and reflections create different approval problems.

Put a one-page campaign brief beside the truth pack. Name the audience, buyer problem, one approved benefit, objective, offer if applicable, CTA, product name, and one visual mood. For example: “Launch [SKU] to [audience]; show [approved benefit]; calm natural-light styling; CTA: Shop [product].” This is not ad copy. It is the constraint that keeps every output recognizably in the same campaign.

How do you set a brand kit that holds seven assets together?

Set the brand kit before you make the hero, since Lamina’s documented workflow scores assets against that kit before distribution. Include logo variants, HEX palette values, heading and body fonts, voice rules, reference images, visual do/don’t rules, and the locked product reference. Put approved CTA language in the kit. Don’t ask the generator to make it up.

Write rules a reviewer can check in seconds. Set lighting direction and softness, background family, crop preference, retouching level, permitted props, forbidden props, and whether people can appear. Add a product rule: “Do not alter label text, logo, package geometry, color, cap, closure, material finish, or product count.”

Deep Banerjee, Head of Business at Lamina, explains why the order matters:

How do you make the hero product image first?

Generate the hero before any lifestyle scene and approve it as the campaign’s visual anchor. Ask for a clean, product-led composition, accurate packaging, a physically plausible contact shadow or reflection, and intentional copy-safe space. Keep props sparse. The hero should make SKU review easy, not bury the product in a decorative set.

Build the prompt around fixed requirements: “Use the uploaded image as the actual product. Preserve silhouette, package geometry, material, exact color, logo, label, and proportions. Create a premium product hero with [approved light treatment], [approved background], and clear negative space for copy. Exclude extra products, altered text, unsupported claims, clutter, and distorted reflections.”

Review the first result beside the truth pack. Check visible label text character by character, then inspect cap shape, seams, edges, finish, and product count. If the hero fails, fix the reference or prompt and regenerate. Never use an attractive, inaccurate candidate as the source for the rest of the campaign.

Which three lifestyle scenes should you generate?

Make three lifestyle scenes that shift context while preserving product truth, rather than three disconnected aesthetics. Keep the approved hero, palette, light family, grade, type treatment, and copy-safe zone fixed. Change one variable per scene: use moment, setting, or level of detail.

Scene one shows the product doing its job in an everyday-use context. A skincare jar can sit on a bathroom shelf beside one approved supporting prop; a drink bottle can appear on a desk in a credible workday moment. Give the product a visible reason to be there.

Scene two serves awareness through an editorial setting that remains inside the brand kit. Keep the color family and lighting logic, then use a more aspirational environment or surface. Scene three supplies proof: an in-hand moment, material detail, or tactile macro that establishes scale and finish. Product context helps shoppers picture use, though Deep Image’s controlled lifestyle-generation guidance warns that contextual generation can change materials and proportions, add components, or produce implausible reflections. Each is a rejection, not a minor retouch request.

Use one prompt structure for all three: product-truth instructions; scene; surface; light; composition; approved palette; one to three supporting props; and exclusions. Change “Scene: morning countertop use” to “Scene: editorial stone shelf” to “Scene: hand holding product at use distance,” while leaving every product-preservation instruction intact.

How do you check product fidelity before making video?

Reject any generated still that changes product truth or introduces an unapproved claim, even if the art direction lands. Lamina’s documented create, track, evaluate, and distribute workflow scores each asset against the brand kit. Put a human SKU and compliance review beside that score.

Use a pass/fail gate before you choose a still for animation. Reject changed logos, unreadable or incorrect label text, the wrong cap or closure, altered color, changed material finish, distorted silhouette, implausible scale, extra components, broken reflections, impossible shadows, duplicate products, and claims the approved brief does not support. Most campaign drift gets caught here.

Record the reason every candidate failed. “Wrong cap geometry” can be acted on; “looks off” cannot. The rejection log improves the next prompt variant and stops a reviewer from approving a later video that carries forward an earlier still’s defect.

What beatsheet works for a 10–15 second product ad?

A short vertical product ad needs to establish the product immediately, prove one thing, and end on one CTA. The master should not attempt every feature or rebuild a full product page. Lamina’s vertical-Reels use case specifies a brand-kit beatsheet of hook, build, payoff, and CTA. Use it so the middle does not turn into unrelated visual effects.

For a 15-second edit, use the opening moment for a reveal or problem cue, then product introduction or interaction, material or feature proof, use payoff, and a close with a clean packshot and approved CTA. In a 10-second version, compress proof and payoff. Keep the final product confirmation.

The master and cutdowns should use the same source still or approved continuity reference. That keeps a bottle’s fill line, a bag’s hardware, or a cosmetic tube’s label consistent between a paid Reel and its retargeting edit.

How should you score campaign consistency?

Score every final asset from zero to two across five checks. Target at least 16 points out of 20, and require full marks for product truth and claim/compliance. A high total cannot excuse a wrong label or invented benefit. Those are hard failures.

Score Product truth on shape, color, material, cap or closure, logo, label, and product count. Score Brand system on palette, typography, grade, visual rules, and CTA wording. Score Scene logic on scale, perspective, contact shadows, reflections, and props. Score Message on the one approved benefit, offer, and CTA. Score Channel readiness on aspect ratio, safe areas, captions, text legibility, and unclipped logos.

Lamina supports rubric-based evaluation against the brand kit before publishing. Use that score as a consistency screen, then leave SKU facts, regulated claims, and final sign-off with the human reviewer.

Which export checks matter for Instagram Reels?

Export Reels as a vertical MP4 or MOV using H.264, with captions and sound, then inspect the actual uploaded preview before publishing. Meta recommends vertical Reels designed at 1440×2560 and asks advertisers to keep critical text and graphics away from top, bottom, and side interface areas.

Keep the opening visual clean. A hook buried under account UI is wasted; a final CTA covered by controls makes a technically valid export commercially useless. Use captions even with voice or music, and keep the product name readable through the final packshot.

For stills, retain the approved high-resolution master, then make placement-specific crops afterward. A lifestyle render may work beautifully for paid social and still be the wrong source for a technical PDP image if the label, dimensions, or physical detail cannot be verified.

What should happen before the campaign is published?

Publish only after all seven files have an approved scorecard, a recorded SKU-truth pass, claim-and-CTA approval, and a placement preview. Human art direction is still required, especially for a brand-critical hero, dense label copy, reflective materials, and regulated product language.

The campaign should read as one idea across seven outputs: one product, one visual system, one proof point, one CTA. If a lifestyle still resembles another brand’s work, or a cutdown makes the product physically different, send it back through the approval gate. Generation supplies volume. Operating discipline makes that volume usable.