Is our product catalog and store data worth anything to AI companies?

Store data can qualify when a product catalog comes with years of orders, returns and reviews tied to the same products. Ten providers name store records or systems as of October 2026.

What do AI companies use store data for?

To train and evaluate models on store operations. idler’s ShelfLife collection uses a decade of an e-commerce company’s operational history and reports model scores. That shows a use of store records; it does not establish a price for another store’s archive.

Marketplace history is the same kind of record. Polyshares’ Amazon guide says “Amazon order history ties SKUs and ASINs to prices, promotions, fulfillment channel and ship-to region over time” (checked September 2026).

Is a product catalog worth anything on its own?

A catalog can qualify, though Polyshares usually sizes a store’s records by orders. Its Shopify guide says “Product catalogs with titles, descriptions, variants, prices and inventory show how merchants describe and stock goods”, but it asks merchants to report total orders and to keep products out of that number “unless the store has few orders and a large catalog” (checked September 2026).

Titles, descriptions, prices and photos are already on your public product pages. What a shopper never sees sits in the export:

• Cost. Shopify’s product CSV has a “Cost per item” column, described as “How much it costs you to make or acquire the product or variant”.
• Stock. Its “Inventory quantity” column applies only to stores with a single location.
• Identifiers and attributes. SKUs, barcodes, product category, vendor and options. WooCommerce’s exporter (Products > All Products > Export) can add custom fields with a meta: prefix (checked October 2026).

Images travel as links. Shopify’s product file lists a “Product image URL”, and WooCommerce’s lists image URLs in its Images column, so the photos themselves are a separate folder to count and a separate rights question (below).

What makes store data worth more or less?

Years of orders, and orders you can follow past checkout. As providers put it, as of October 2026:

What adds value	In the provider’s words
Years of order history	Polyshares counts a store’s years from “the date of the store’s first order”; Troveo’s “Fashion E-commerce Platform” snapshot lists 10+ years
What happened after the order	Miro Advisory’s e-commerce workflow runs from “Order through fulfillment, support, return or retention”
Orders tied to the money and the work	Avelence’s finance and operations card, which names Shopify, puts orders beside invoices and reconciliation history

Tickets about orders have their own page. Rights in photos and copy, and customers’ details, can narrow the scope; both are covered below.

Which buyers say they want store data?

Ten providers meet this test as of October 2026. Across the 25 providers Data Licensing Report tracks, their own pages must name product catalogs, store orders, retail purchase histories, a store platform, or operational data taken from an online store.

Provider	What it is	What its own pages name
Avelence	Introducer; does not buy data	Shopify, with QuickBooks and NetSuite, on its finance and operations card
Corpus	Licenses directly	Shopify among its “Customer, sales & finance” systems
FileYield	Listing marketplace	Menus & Product Catalogs in its document catalog, plus purchase histories in retail
idler	Licenses directly	ShelfLife, built from an e-commerce company’s records
Mercor	Licenses directly	Retail examples: “Order records, inventory logs, listings”
Miro Advisory	Introducer; does not buy data	“e-commerce and fulfillment” among the systems it lists
Polyshares	Licenses directly	An Ecommerce group, “Orders, products and customers”, with guides for Shopify, Amazon Seller Central, WooCommerce and BigCommerce
Sell My Business Data	Introducer	Orders and customer history, with Shopify among the systems it names
Telegraph Lab (affiliated with Data Licensing Report)	Licenses directly	Shopify under “Operating records”; “Retail & ecommerce” among target industries
Troveo	Licensing agent	Shopify among its finance and operations systems

Replay and Scale AI name e-commerce as an industry but do not identify store records or platforms on those pages, so they are outside this record-specific table. The e-commerce page distinguishes those industry matches. Ask a provider which tables it wants and which customer fields it removes.

Which platform holds our order history, and since when?

On the platform where each order was placed, and each platform hands it over differently. As of October 2026:

Platform	Where the history sits	How it comes out
Shopify	In Shopify, from the store’s first order	Admin CSV exports of orders and products; the order file’s personal columns and Shopify’s limits on apps are on the e-commerce page
WooCommerce	In your own WordPress database; since version 8.2 (October 2023), new stores keep orders and addresses in dedicated order tables	The built-in exporter handles products; orders stay in the database
Amazon Seller Central	With Amazon	The all-orders report “Contains all orders that were placed in the specified period”, at most 30 days per request for these order-tracking reports
Reviews	In the platform or a reviews app	That app’s export
Product photos	Platform storage, a photographer’s delivery folder or a shared drive	As files; exports list them only as links

A store that changed platforms may keep its early years only in the old one’s exports (old systems).

Can we license data we downloaded from Amazon or Shopify?

The platform’s seller terms govern part of the answer, and the two read differently.

Amazon. Amazon’s agreement-update announcement for March 4, 2026, published by News_Amazon, says it added restrictions on using Amazon materials or services for AI development. The announcement also says the US/Canada and Mexico agreements are separate. Compare the files a buyer requests with the Business Solutions Agreement that applies to your store.

Shopify. Its Terms of Service (updated August 1, 2026) define a merchant’s Materials to include the products you sell “(including description and price)”, photos and “product reviews”, and section 8.1.3 says Materials you owned before uploading them “remain yours, subject to any rights or licenses granted in the Terms of Service or elsewhere”. Apps and connectors are bound by Shopify’s separate API terms, which the e-commerce page quotes.

Whose rights are in our product photos and descriptions?

Under US law, copyright starts with the author. For a work made for hire, the employer or commissioning party is treated as the author unless a written agreement says otherwise (17 U.S.C. §201). A copyright-ownership transfer generally needs a signed writing (§204); the job photos page explains the distinction. In a catalog, also check:

• Supplier and brand content. Images and descriptions a manufacturer supplied so you could sell its products, and the terms that came with them.
• Copied text. Specifications or descriptions taken from a manufacturer’s site.
• Freelance product shoots. What the photographer’s contract says about ownership or license.

Shopify’s clause above speaks only to Materials you owned before uploading them. Note in your inventory which photos and copy your staff produced, which came from suppliers and which from freelancers.

What about customers’ details in order histories and reviews?

Some providers exclude identifiable customer data from the delivered package. Corpus lists “Customer PII in identifiable form” as out of scope. Polyshares says Amazon’s All Orders report “does not include customer-identifying information”, and Amazon’s data protection policy for its seller API says developers “will retain PII for no longer than 30 days after order delivery” with a longer period only when legally required (Amazon’s developer guide, checked October 2026).

Shopify’s order export includes names, email and addresses. WooCommerce’s order tables include order addresses. Review exports can contain a reviewer’s name and text, so count them separately. Whether licensing identifiable orders is a “sale” under California law, and what your privacy policy promised, are on the e-commerce page; what “anonymized” means covers what has to change before order files can go.

What could our store data be worth?

A store-specific offer needs an inventory; the public examples above do not establish a rate per order or product. Troveo’s Fashion E-commerce Platform snapshot gives years and systems without a price. Polyshares’ deal examples, Corpus’s payout examples and Telegraph Lab’s published deal cards describe other businesses.

For figures by company size, see How much is my company’s data worth?; the e-commerce page gives the one online-retailer transaction Telegraph Lab has described to Data Licensing Report, which does not say which records went. For a buyer, describe your store by orders per year, the first order’s date, the platforms, and whether customer fields, reviews and photos are in or out.