From 11,400 supplier rows to a live catalog in nine days

Nordkust Supply is a Nordic outdoor and workwear distributor with four suppliers, three languages of product data, and a catalog that had stopped growing for one reason: nobody had time to build the products.

Before Loger, one person handled product setup. Every new supplier file meant reading a spreadsheet with shifting column names, hunting through an image folder with SKU-based file names, splitting sizes and colors into variants by hand, and writing titles and descriptions from scratch. The team measured it once: five to six minutes per product, on a good day.

The starting point

The first import was a single supplier: 11,400 rows across seven sheets, 9,800 images in nested folders, and no consistent naming between the two. Roughly 2,600 real products were hiding in there once variants were grouped correctly.

At six minutes per product, building that by hand would have taken about 260 hours. That is more than six full working weeks for one person, on a catalog that would need updating again the following quarter.

What the import actually did

The spreadsheet and the image folder went in together. Loger read all seven sheets in one pass, detected which columns held SKUs, prices, sizes, and colors, then reconstructed 2,614 products with proper option sets instead of 11,400 flat rows.

Images were matched to products and to individual variants, including folders where the file names only partially matched the SKU. Around 4% of images landed on the wrong variant and were corrected in review — 380 fixes instead of 9,800 manual attachments.

Titles, descriptions, and metadata were generated in bulk in the tone the team had already defined, then read through before anything was published.

Review, not rebuild

The important shift was where the time went. Instead of building products, the team reviewed them. The review workspace flagged 212 products with missing dimensions, 96 with duplicated descriptions, and 41 that were duplicates of an existing catalog item.

Those 349 exceptions were the actual work. Everything else was already correct and simply needed a second pair of eyes.

The numbers

Nine working days from first upload to a fully published catalog, against an estimated six and a half weeks by hand. Around 2,614 products live, roughly 232 hours saved, and a first-pass accuracy of about 87% before any manual correction.

The second supplier took two days. The third took less than one, because the pricing rules, tone, and collection structure were already in place and every later import inherited them.

What changed after

The catalog stopped being a quarterly project. Price and stock updates now run from a fresh supplier spreadsheet matched against the live catalog, so only the differences get applied. Seasonal collections are queued and scheduled instead of published in a rush.

The person who used to build products full time now owns merchandising and supplier relationships. That is the part of this study worth copying: the work did not disappear, it moved to something that actually compounds.

How to repeat it

Start with one supplier file end to end rather than everything at once. Keep image folders named after the SKU or product name. Leave variants flat in the source and let them be rebuilt. Define your tone before generating copy in bulk. Then publish in batches so every change stays measurable.