A product banner can become false between approval and impression. The price changes. A color sells out. The product image points at a discontinued variant. A sale badge remains after the promotion ends. The creative file may be visually perfect and commercially wrong.
Feed-driven production reduces manual copying, but it does not remove the need for governance. Bannerify can use CSV or XLSX data to swap copy, images, media, styles, and visibility while exporting variants as HTML, MP4, GIF, WebM, APNG, WebP, images, or PDF. Ecommerce and media teams still need to define which catalog fields are trustworthy, how they map to creative, and what happens when data is missing or stale.
Define the campaign’s data boundary
Do not connect the entire commerce catalog to a banner template. Create a campaign-specific export containing only approved products and fields.
A useful row might include:
| Field | Example | Control question |
|---|---|---|
| campaign ID | spring-running | which media order owns this row? |
| product ID | SHOE-1042-BLK | does this identify the exact advertised variant? |
| product name | Trail Runner 4 | what is the approved display length? |
| current price | $129 | does tax treatment match the market? |
| previous price | $159 | is comparison pricing legally supportable? |
| image URL or asset | approved packshot | does it match the SKU and color? |
| destination URL | product route | does it retain required tracking? |
| availability | in stock | how fresh must this value be? |
| offer end | timestamp | which timezone governs shutdown? |
The feed should not contain customer or behavioral data. Audience selection belongs in the advertising platform; creative production needs controlled product facts.
This article is different from the spreadsheet-driven banner variant workflow, which covers general campaign permutations. Here the hard problem is catalog truth: matching product identity, price, stock, imagery, and destinations at the moment creative is generated and served.
Set field authority and freshness rules
Name one source for each fact. Product marketing may own the short name, commerce owns price and availability, brand owns image selection, legal owns offer qualifiers, and media operations owns destination tracking.
Define refresh expectations before production:
- evergreen name and image data may be reviewed per campaign;
- prices may require a fresh export immediately before trafficking;
- availability may be too volatile for baked-in creative;
- sale end times require an explicit timezone;
- discontinued or suppressed SKUs must fail closed.
If the exported banner cannot update after trafficking, do not present it as dynamically current. Either regenerate it on a safe cadence or use language that remains true, such as “Explore the collection,” instead of a fragile stock claim.
Design a template that tolerates real catalog data
Create fixtures for short and long product names, large and small prices, portrait and landscape packshots, products without a previous price, and missing optional badges.
Define:
- maximum lines and minimum type size for names;
- image fit and safe crop behavior;
- price hierarchy and locale format;
- rules for showing or hiding comparison prices;
- badge priority when several promotions apply;
- CTA wording and destination behavior;
- background contrast across light and dark imagery;
- a safe fallback when any required field fails.
Do not automatically shrink text until it fits. An unreadable product name technically fits but fails the ad. Create alternate layout variants for known content classes, or reject rows outside the approved limits.
Animation should reveal information in a useful sequence: product, reason to care, offer if valid, and action. Avoid placing the only price or qualifier in a frame that disappears too quickly. The animation timing guidelines help separate readable pacing from decorative motion.
Validate the feed before generating files
Run checks before importing the CSV or XLSX:
- unique campaign and variant identifiers;
- required values present;
- known market and currency codes;
- parseable dates with timezone;
- approved product IDs only;
- image files or URLs reachable and correct;
- destination URLs on allowed domains;
- comparison price greater than current price when shown;
- no HTML, spreadsheet formulas, or unexpected control characters in display text;
- no expired offers or suppressed inventory.
Keep a rejection report. Silently substituting blank copy or a default image can create a plausible-looking but incorrect ad.
Generate a manifest with the creative
Use Bannerify to apply the approved data to the design set and export the required sizes and formats. Produce a manifest beside the files containing:
- source row and product ID;
- creative size and format;
- filename;
- destination URL;
- offer and expiry;
- generation timestamp;
- template version;
- review status.
Use stable filenames such as campaign-product-market-size-version. Avoid relying on final or the row number of a spreadsheet that will be sorted later.
For HTML5 placements, verify click tags, platform requirements, and package size using the trafficking handoff checklist. A correct catalog row does not guarantee an acceptable ad package.
Review the highest-risk combinations
Do not approve only the most attractive products. Review extremes:
- longest product name with the smallest banner;
- light packshot on a light background;
- four-digit localized price;
- absent previous price and badge;
- a product near offer expiry;
- a destination with tracking parameters;
- fallback creative after a required-field failure;
- static fallback for an animated HTML placement.
Check every unique layout rule and a sample from each market. Automated screenshots or a Bannerify preview page can make review faster, but the approver needs the manifest and source facts beside the creative.
Plan shutdown before launch
Every offer-led batch needs an owner, expiry, and removal path. Decide whether the media platform, campaign schedule, or operations checklist stops delivery. Record what replaces expired creative. If availability is important but cannot be refreshed in the ad, route the destination to a useful category page or configure a safe fallback at the commerce layer.
After launch, sample live placements and click through them. Confirm that the displayed product, price, destination, and landing-page offer still agree. Monitor rejected ads and broken destinations separately from design approvals.
Release checklist
Before trafficking, confirm that:
- every field has an authoritative owner;
- the feed contains approved products only and no customer data;
- freshness and shutdown rules match the volatility of each claim;
- product images, names, prices, variants, and URLs refer to the same item;
- templates handle long, missing, and localized content safely;
- invalid rows fail visibly instead of producing plausible defaults;
- exported files and manifest share stable identifiers;
- file size, click tags, animation, and fallback assets meet platform rules;
- live placement and landing-page checks have named owners;
- expired and unavailable products have a safe replacement path.
Feed-driven banners should reduce copying, not automate mistakes. Keep catalog truth constrained, make failure obvious, and use Bannerify to generate controlled variants from a template that has already survived the messiest products in the feed.
Read as Markdown Related resource index Site index for AI agents
Pro Bundle