Every supplier sends a different spreadsheet

Ask anyone who runs a multi-brand store what takes the most time, and you'll get a version of the same sentence.

Adding a new supplier costs us a week.

Not because the products are hard. Because the file is. One brand sends Excel with merged header rows. One sends a CSV where price is 24,90 and another where it's 2490. One sends an XML feed that changes structure whenever their developer feels like it. One sends a link to a Dropbox folder of photos named IMG_2231.jpg.

There is no standard. There never was. Every supplier built their export around their own warehouse system, and none of them built it around you.

Why the files will never match

It's tempting to think this is a discipline problem - that suppliers should just agree on a format. They won't, and it isn't unreasonable of them.

A supplier's file is an export from whatever runs their business. The column names are their internal field names. The units are their internal units. The categories are their internal categories, which exist to organise a warehouse, not to sell to your customers.

Retail standards exist. Almost nobody sends you one. What you get is the export button.

So the work lands on you: take fifty shapes and make them one shape. That translation is the actual job of running a multi-brand catalogue, and it's almost always done by hand, in a spreadsheet, by the one person who knows how.

The three ways supplier files break quietly

Loud breakage is fine. A file that fails to import gets fixed by lunchtime. The expensive failures are the ones that succeed.

1. Column drift. A supplier renames EAN to EAN_Code. Your import still runs. It just stops bringing barcodes. Nothing errors. Three weeks later somebody notices that new products from that brand are disapproved on Google, and nobody can say when it started.

2. Format mismatch. Price arrives as a string with a comma, and lands as zero. Weight arrives in grams into a kilogram field. Sizes arrive as M where the rest of your catalogue uses Medium, so variants split and your filters stop working. Every one of these imports cleanly.

3. Identity collisions. Two suppliers use the same SKU for different products. Or the same product arrives from two suppliers under two SKUs, and now you have duplicates competing with each other in ads, splitting their own performance.

The pattern is the same each time: the system said yes, the data was wrong, and nothing told you.

What a week of onboarding actually goes on

If you break down that week, very little of it is thinking:

• Opening the file and working out which column is which
• Renaming and reordering columns to match your store's import template
• Fixing decimal separators, units, and category names
• Chasing the supplier for missing barcodes and images
• Rewriting descriptions that arrived as three words or as a wall of the supplier's own marketing
• Uploading, finding errors, fixing, uploading again

Then next month the supplier sends an updated file, and roughly half of that happens again - because the mapping lived in someone's head and in a spreadsheet that got overwritten.

That's the real cost. Not the week. The fact that the week doesn't accumulate into anything.

What "map it once" looks like

The alternative isn't a better spreadsheet. It's making the mapping a permanent object rather than an afternoon's work.

A template per supplier. You match their file shape to your catalogue once - their Artikelnummer is your SKU, their Farbe is your colour, their price column is in euros with a comma. Saved. Named. Reusable.

Change detection instead of silent drops. Next time the file arrives with a renamed or missing column, the template flags it and asks. It does not quietly import a catalogue with no barcodes.

Transformations, not manual edits. Comma to point, grams to kilograms, M to Medium, category names to your taxonomy - as rules that run every time, not edits somebody remembers to make.

Per-product exceptions. This is the part most tools miss. Real catalogues don't obey one rule. One brand really does put the size in the title. One range really has no barcodes. A global template applied to everything breaks on every exception, and there are always exceptions. The mapping has to allow "except these forty products", or you'll be back in the spreadsheet within a month.

One catalogue underneath. Once supplier files land in a single catalogue, everything downstream reads from that one place - your store, your Google feed, your Meta catalogue, your marketplace listings. Change a price once. It's a smaller idea than it sounds and it removes an entire category of Monday morning.

Before you buy any tool for this

Four things worth doing first, whether or not software is involved:

• List your suppliers by pain, not by revenue. The brand that takes three days a month to maintain may be smaller than the one that takes an hour. Fix by time spent.
• Write down one target shape. What are your canonical fields, units, and category names? Until that exists on paper, every mapping is a negotiation.
• Decide what "complete" means for you. A product with no barcode and no image can enter your store, but it can't enter Google or Meta. Knowing which state each product is in matters more than filling every field.
• Ask suppliers for one thing, not ten. You will not get them to change their export. You may well get them to add a barcode column, or a link to a photo folder. Ask for the single field that unblocks the most products.

The version of this that scales

Feedwrk pulls supplier files - Shopify store, XML, CSV - into one catalogue, keeps the mapping per supplier, flags a file that changed shape, and pushes finished products out to the channels you sell on. Products that aren't ready are visible as not ready, per channel, with the reason.

Shopify stores first. Sources are Shopify, XML and CSV today; more store platforms are next.

If your catalogue already lives in a Shopify store, start with the scan. It's the fastest way to see which suppliers are costing you the most products.