Licensing checklists

Product catalog licensing: field sources and variant checks

Separate product text, specifications, and images by source; document units, variants, snapshot dates, and matching-test limitations.

Catalog exports often mix original descriptions, supplied specifications, and third-party imagery. A coherent inventory separates those components and explains relationships between products and variants. The goal here is to define a reviewable collection and a delivery plan before discussing commercial terms.

Build the source register at field level

Catalog ownership is rarely explained by a single label. Record who supplied descriptions, specifications, taxonomy, and images, plus the source record for each group. A product row may contain both internally written text and supplier-provided content. Keep unresolved components visible and separable for review rather than calling the entire export internally authored.

Distinguish products, variants, and listings

Define what each identifier represents. A product family, size/color variant, seller listing, and historical price observation are different units. Record the keys connecting them and the handling of merged or retired IDs. Count each level separately; duplicated listings should not be reported as distinct product coverage without explanation.

Make units and missing values explicit

Document original and normalized units for numeric attributes. Distinguish unknown from not applicable and deliberately removed values. Record transformations such as decimal parsing or unit conversion and identify fields that remain free text. Review a sample of extremes and mixed-unit records rather than relying only on successful file parsing.

Choose a snapshot or history deliberately

Give the export a timestamp and describe whether discontinued items remain. Historical prices, changed specifications, and current product descriptions need clear temporal relationships. Ask the evaluator which periods matter before preparing repeated snapshots. A promise of ongoing freshness creates work beyond a one-time delivery and needs separate discussion.

Design a matching evaluation around difficult pairs

If the proposed task is product matching, include examples that distinguish near-identical variants and examples that share a name but differ in specifications. Group related product families when designing separate evaluation subsets. Have the evaluator define success criteria; a well-structured catalog alone does not demonstrate model performance.

Catalog review sheet

Edit this outline, then copy or download it. It stays in this page and is not submitted or saved by this tool. Keep confidential records and personal information out. Save your work before leaving or reloading.

You can select and copy the text manually. This is a preparation worksheet, not a legal agreement.

Sources and scope of this guidance

Sources checked October 19, 2026. AWS describes field dictionaries as an aid to evaluating a dataset. The workflow and examples here are our preparation suggestions, not AWS requirements or a claim of platform eligibility.

AWS Data Exchange: data dictionaries