From Excel to PIM: how to migrate your catalog without losing data

The spreadsheet is not the problem

Almost every catalog starts in Excel, and that makes sense: it is flexible, everyone knows how to use it and it costs nothing. The problem comes when the catalog grows: there is one version per channel, another from the supplier, another someone edited on their laptop, and nobody knows which is right.

Moving to a PIM does not mean throwing that work away. It means the spreadsheet stops being the source of truth and becomes what it is: a way to load and download data.

The only must-have: a unique SKU

Before importing, check one thing carefully: that every product has a unique SKU that does not change. The SKU is the key that links information across systems; if two products share a SKU, one will overwrite the other, and if a product changes SKU between files, it will appear twice.

Everything else —different column names, incomplete data, values written in several ways— is handled during the import or afterwards.

Prepare the file

A good import file does not have to be pretty, it has to be readable:

• One row per product and one column per field. No merged cells and no titles in the middle of the table.
• Format: UTF-8 CSV or XLSX. Encoding matters: a CSV saved in another encoding turns accented letters into strange symbols.
• Clear separators: if a cell holds several values, such as “wood, metal”, always use the same separator.
• Variants: if you sell sizes or colors, a column with the parent product's SKU lets you build the groups in the same load.

The first load: simulate before writing

The difference between a calm migration and one that leaves the catalog in a mess is reviewing before writing. A good import tells you:

• which column fed each field (name, brand, category);
• which columns can become filterable attributes and which cannot, and why;
• how many products will be created and how many updated.

With that simulation you decide what to keep as an attribute, what to leave as source data and what to fix in the file before continuing.

What to do with the extra columns

An ERP catalog brings many columns that are not attributes: internal codes, notes, raw descriptions. Don't throw them away. Kept as source data, they still help generate descriptions and keep information from being lost, even if they cannot be filtered in the store.

Columns that do describe the product —color, material, dimensions— are worth turning into attributes from the first load.

After the import

With the catalog inside, the work shifts from “gathering data” to “completing product records”:

1. Measure what is missing for each product according to its family and each channel.
2. Complete the best sellers first.
3. Generate the missing content from the real data.
4. Export to each channel from the PIM, and stop editing in spreadsheets.

The last point is the hardest and the most important: if people keep fixing things in Excel, in two months there will be several versions again. Agree with the team on a date after which the spreadsheet is only used to load data, never to correct it.

How Bruxel handles it

Bruxel imports UTF-8 CSV and XLSX —CSV up to 256 MB and XLSX up to 10 MB, with up to 50,000 rows— and recognizes several names for key columns, such as sku, ean, name or brand. Before writing, “Review before importing” shows what it would do with each column; columns that match an attribute fill in by themselves, and a later import never overwrites a value someone corrected by hand. The guide to importing your catalog shows it step by step, and the first day use case walks from import to download.