Embedded product-list JSON

Listing pages have to render fast, so retailers print the shelf data into the first response. On Walgreens.com - Pharmacy & Drug Catalog (scrapeable), those server-rendered category pages carry an embedded productList holding, per the cataloged record, product names, regular and sale prices, unit prices, UPC/GTIN, ratings, inventory status and ingredient text for roughly 72 items per listing page.

The dataset's declared shape matches the mechanism: data_formats ['html','embedded_json'], granularity 'SKU/product level with per-listing-page batches of about 72 items', tens of thousands of OTC, wellness, beauty and household SKUs across paginated categories. GoodRx appears in the same drug-retail slice with 'embedded_json' and 'json_ld' among its formats, so the pattern is not unique to one pharmacy chain.

Competitive price and assortment intelligence in drug retail needs exactly the fields productList carries - and needs them tied to identifiers that survive joins.

• UPC/GTIN presence makes cross-retailer matching possible. Price comparisons keyed on product names alone collapse on pack-size differences; unit prices plus GTINs hold the comparison honest.
• Inventory status turns price data into availability data. A cheaper competitor price means nothing if the item is out of stock at every store.
• Batch structure sizes the crawl. At ~72 items per listing page across paginated categories, page-count arithmetic estimates corpus size before you start.
• The counterweight is terms: Walgreens' recorded terms permit personal non-commercial shopping use and expressly prohibit automated collection methods such as robots, crawlers and scraping mechanisms.

1. Confirm the batch constant. Parse several category pages and verify the ~72-items-per-page figure holds across categories - drift signals either changed pagination or truncated payloads.
2. Validate identifier completeness. Check that UPC/GTIN fields populate on every record, not just featured items; identifier gaps decide whether cross-source joins work at all.
3. Test price-field pairing. Regular and sale prices plus unit prices must co-vary sensibly (sale <= regular; unit price = price / size) - inconsistencies reveal stale cache layers.
4. Exercise inventory-status semantics. Record what statuses appear and how they track real stock; Walgreens updates daily in the catalog, so status churn should be visible week to week.
5. Weigh terms before scheduling. The cataloged terms prohibit automated collection outright - evaluate the technical exposure and the permission question as separate gates.

Where to see it in context: drug-retail data.

• national-drug-code - the FDA identifier system that complements UPC/GTIN when matching pharmacy products across sources.
• store-inventory-availability - the stock-status dimension exposed as a field in these embedded lists, tying each SKU to availability.