Amazon Is Not a Source: Mirroring Product Data to OTTO Fails

If you sell on Amazon, you have already prepared your range: titles, images, attributes, variants. The obvious conclusion is to simply transfer these listings to OTTO. That conclusion is wrong, and not because a connector is missing, but because of the data logic. Amazon is a sales channel, not a source system. If you use it as a source anyway, you carry its quirks into every further channel.

Two marketplaces, two data models

On Amazon, the product type determines which data is required. The Product Type Definitions API provides a JSON schema that describes requirements, attributes and the conditions under which a requirement applies. OTTO works with its own category structure: the categories are retrieved through the partner interface, and the category names in your data must match the exact names used by OTTO. Each category brings its own attributes. Attributes classified as LEGAL are mandatory, and attributes with high relevance improve visibility. Brand IDs add another layer: the brand of your products must match the brand IDs used by OTTO, and EAN, SKU, category and brand ID are required fields when you create a product.

An Amazon product type therefore cannot be translated one to one into an OTTO category. What is complete on Amazon may lack a mandatory attribute at OTTO. Conversely, OTTO asks for values that were never maintained in your Amazon listing because Amazon did not need them.

Images, titles and variants: where copying breaks

Images

Both platforms want a main image on a white background, but not the same image. For the main image, Amazon states pure white (RGB 255, 255, 255), at least 1000 pixels on the longest side and a product share of 85 to 100 percent of the frame. OTTO requires a cut-out on white with a continuous white border on all four sides, a product share of 85 to 95 percent of the image area, a short side of at least 500 and a long side of at least 1000 pixels, a maximum edge length of 4500 pixels and no more than ten images per variant. Without at least one image per variant, a product does not go live at OTTO.

Titles

Title rules are not a constant. For Amazon.de, Amazon announced an upper limit of 200 characters for most categories (effective since 21 January 2025), plus a list of disallowed characters and the rule that a word may appear at most twice (prepositions, articles and conjunctions are excepted). For Amazon.com, Amazon announced a limit of 75 characters for all categories except media, to apply from 27 July 2026. If you copy titles, you copy the state of a rule that differs by marketplace and point in time. The public OTTO documentation contains no comparable title rule; check the requirements of your category in the partner area.

Variants

Amazon describes variations through the parentage level, the relationship between parent and child SKU, and a variation theme. OTTO starts from the orderable unit: the variation (SKU) is the basic entity, product data is managed primarily at variation level, and a product groups variations that share brand, category, name and dimensioning. Which attributes may form variants is defined by the variation themes of the respective OTTO category. An Amazon parent-child structure therefore does not automatically carry over to OTTO.

Keeping it in sync is the real effort

The initial transfer is the smaller part. At OTTO, products are processed asynchronously: the response 202 only confirms receipt. Only the update task shows whether the data was accepted, and a second validation step on otto.de reports its errors separately in the marketplace status. Data without changes is discarded as unchanged, even if it failed before. Stock updates only work for SKUs that already have a MOIN, and OTTO recommends retrieving category and brand lists regularly, ideally daily. On the Amazon side, listings are maintained per SKU through the Listings Items API, and problems appear as issues on the listing.

If you mirror from Amazon to OTTO, you end up with two worlds of errors that nobody sees in one place. Every correction on one side has to be carried over to the other by hand, and every new obligation of one marketplace catches the data of the other.

PIM with AI as the source instead of Amazon

The solution is to reverse the direction. Amazon does not feed OTTO; a PIM with AI feeds both. In the Golden Record, ainavio merges the product data from all sources into one record per product, and the sources can be prioritized. The inheritance logic automatically transfers shared properties to related variants, so they do not have to be maintained again for each variant.

The process in practice:

• Source: The Golden Record in the PIM is the single truth, not the Amazon listing.
• Mapping: The Marketplace Agent automatically assigns categories and mandatory attributes to the fields of 26 shops and marketplaces, including Amazon, OTTO and Kaufland.
• Enrichment: The Enrichment Agent adds missing attributes and creates titles, descriptions and bullet points, so the mandatory fields of the second marketplace are filled as well.
• Images: The DAM adapts images to the channel with AI image editing: background removal, cropping, cut-outs and channel-specific image adjustment.
• Check: ainavio checks completeness, consistency and channel conformity before every export. The Quality Agent detects missing information, duplicates and inconsistencies.
• Delivery: The integrations for Amazon and OTTO keep changes to images and descriptions up to date through real-time synchronization, and prices and content can be controlled centrally per channel.

This does not replace the platform rules, it makes them manageable. A new title limit or an additional mandatory attribute is stored once as a channel rule instead of being corrected in hundreds of listings.