Health and drugstore. Shopify for large ranges across many channels.

A brand store sells a few hundred of its own products. An online drugstore carries tens of thousands of products from many manufacturers, often with similar names, different pack sizes and variants. Every product needs an unambiguous identity, otherwise reviews, stock or marketplace offers end up on the wrong product. The problem is rarely the platform, but the quality and matching of the data.

On top of that, product data comes from many sources: manufacturers, wholesalers, the ERP and legacy store systems. Each source has its own fields and spellings. If you merge them without checks, you get duplicate products, empty categories and filters that find nothing. With a range of this size, that does not show on a single product, only in the sum.

That is why we start with the data model: which identifier leads, which system supplies which data and what happens when information is missing. At Kanela, 337,891 product-category relationships were analysed and 13,540 category tags added. Only on that basis were navigation, filters, search and bestseller logic designed around the large range.

In a drugstore a product rarely has just one number. There is the SKU in the store, the manufacturer's EAN or GTIN, in Switzerland often a Pharmacode, plus legacy product IDs and separate identifiers for each marketplace. As long as these numbers are not cleanly linked, errors appear at every interface, and nobody can see at once where they come from.

At Kanela that was exactly the core of the starting point. Product identities across SKU, EAN/GTIN, Pharmacode, legacy product ID and marketplace ID were unified. For Manor, existing offers were matched to Shopify variants by GTIN, 3,159 offers in total. For Galaxus, the case study lists 32,203 consistent identifiers that are validated separately before being supplied.

Marketplaces bring reach, but each has its own formats, rhythms and rules for product data, prices, stock, orders and cancellations. If you send everything through a single feed, you often only see errors once customers have already ordered. That is why we separate every data flow and check it before it reaches a marketplace.

At Kanela three marketplaces run with separate workflows. Zur Rose receives daily product data plus price and stock updates every 15 minutes, with 27,302 products validated there. Galaxus gets five separate feeds and a protected openTRANS workflow. At Manor, existing Mirakl identities remain intact while offers, prices and stock move to the new workflow in stages.

The opposite direction matters just as much. Marketplace orders have to arrive in the system exactly once, and shipping and cancellations have to be reported back. An explicit channel source, duplicate protection, a global job lock and a permanent shipping journal prevent duplicate processing. That keeps every transaction traceable months later, when customers or marketplaces ask.

Many drugstore and healthcare stores do not only sell what sits in their own warehouse. Part of the range can be delivered at short notice from wholesalers. The store therefore has to decide from several sources whether a product can be sold. If a message is missing or arrives late, sales stop for no reason or you sell what is not there.

At Kanela a dedicated application combines own stock with availability at Galexis and Avosana. Missing or stale data therefore cannot trigger false sales stops. The application automatically stops unusual bulk changes, backed by dry runs and safety thresholds. In the project, 1,820 Shopify inventory values were also corrected before the new logic took over.

Over the years a large range collects many addresses: products, categories, manufacturers and guides, plus a separate version for each language. When you change platform, almost all of them change. Without complete redirects, search engines lose the mapping, and links from comparison sites, guides or old newsletters lead nowhere.

At Kanela, 147,753 legacy URL mappings from four language versions were created and then read back. Reading them back is the decisive step, because it checks whether every redirect has actually arrived in the store. In parallel, products, categories, customers, bonus balances, orders and reviews were systematically reconciled and transferred.

In Switzerland, several languages are not an extra feature but everyday business. German, French and Italian are national languages, and customers expect product pages in their language. With Shopify Markets and translated content, language versions run in one store. What matters is that every language keeps its own stable URLs and that redirects are planned per language.

Health-related claims are tightly regulated. For food and food supplements, the EU Health Claims Regulation applies and only permits authorised claims. For medicines and medical devices, the German Heilmittelwerbegesetz, among other laws, governs what may be advertised, and Switzerland has rules of its own. The legal review belongs in the hands of your specialists.

The store has to reflect the result technically. Mandatory information belongs in structured fields, not in free text. For food supplements that means ingredients, recommended daily intake and warnings, for cosmetics the ingredients under the Cosmetics Regulation (EC) 1223/2009. Maintained as metafields, they appear in the same place on every product page and can be passed on to marketplaces.

That also helps with maintenance. When a piece of information changes, it is corrected in one place instead of in thousands of descriptions. Claims in banners, category texts and campaigns should go through the same approval as product texts before publication, so that no claim reaches the store or a marketplace without being checked.