Lulu Hypermarket Data Scraping

Six storefronts, six assortments, and one brand name

Lulu operates online storefronts across the UAE, Saudi Arabia, Qatar, Kuwait, Oman and Bahrain. They share a brand and almost nothing else that matters to a dataset.

• Assortment differs materially. Sourcing, import regulation and local demand mean a product ranged in the UAE may not exist in Oman.
• Pricing is set per country, and the gap is not explained by FX.
• Promotional calendars differ, and national events do not align.
• Pack sizes differ for the same brand, so a per-unit comparison needs parsing rather than a price ratio.

Every record carries country and storefront. Cross-country comparison is a deliberate join with FX stamped per observation, not something we do for you by averaging.

Where a client wants a regional view, we deliver it as a computed rollup with the country records intact underneath, so the number can always be decomposed back to what it came from.

EAN is the field that makes the whole thing usable

The inquiry behind this page listed EAN alongside name, brand, size, price, promo price and availability. That was the right instinct, and EAN is the one doing the heavy lifting.

• Cross-retailer joins — Lulu against Carrefour against a marketplace — only work reliably on a shared identifier. Name matching across Arabic and English listings does not.
• Cross-country joins for the same product depend on it too.
• Your own catalogue join is the point of the exercise for most brands.

The honest part

EAN is not always published. Where it is absent we do not invent one and we do not infer it from a similar product. The field is null with a reason, and we report the EAN fill rate per category in the pilot.

That number decides whether a cross-retailer programme is viable in your categories. A category where EAN populates on 40% of listings needs a different approach — attribute and image matching, with a confidence score — and you should know that before commissioning rather than after.

Pack architecture and unit price in a multi-origin market

GCC grocery carries products sourced from Europe, Asia, the Americas and locally, and pack sizes reflect their origin rather than a local standard. The same brand appears in 500g, 1lb and 450g formats across storefronts.

• Unit price is the only comparable figure across those.
• Parsing pack from the title is where most feeds fail, because the format varies by origin and by language.
• Multipacks and bundles need the unit count separated from the pack size.

We parse pack_size, pack_unit and unit_count, compute price_per_unit on a stated basis, and flag pack_parse_confidence where the title was ambiguous. Where we cannot parse it confidently, the unit price is null with a reason rather than a number derived from a guess.