Product Data in Excel: From Chaos to Catalogue

It is Monday morning and you open productlist_FINAL_v7_copy.xlsx. Somewhere along the way, column H has turned the EAN codes into scientific notation (4.71E+12). The same office chair appears three times: once in anthracite, once in dark grey, once in grey/black, because three suppliers each spell it differently. And the new price list that has just landed in your inbox needs to be retyped by hand into your webshop, and then once more into the marketplace.

It works. Until it does not. At a thousand products, Excel is a perfectly good scratchpad. At fifty thousand, it is a time bomb. And here is the thing: product data in Excel does not break with a loud crash. It breaks with silent errors that you only see months later, in your returns, your conversion and your position in Google.

Why everyone starts with product data in Excel

Excel is the cheapest, fastest and most forgiving way to get started. Anyone can use it, it costs nothing extra, and within five minutes you have a tab with your first two hundred items. That is not a foolish choice. It is the logical one. The problem is not that you start in Excel, the problem is that you stay stuck in it while your range grows out from under you.

And Excel breaks more quietly than you think. A review of 35.5 years of research led by Professor Pak-Lok Poon (Frontiers of Computer Science, 2024) found that 94% of business spreadsheets used in decision-making contain errors. Ray Panko (University of Hawaii) measured an average error rate of around 5% of formula cells. So those errors are almost always in there. You just cannot see them.

Where it breaks: five things Excel cannot do

A spreadsheet is a grid of cells. Nothing more. The moment you turn it into a system of record, it lacks exactly the things a catalogue at scale needs:

• No golden record. Three suppliers, three rows for the same chair. Excel has no idea it is one product.
• No dedup. Matching on EAN, SKU or barcode is a manual job, and purple, lilac and violet stay three separate colours.
• No per-field provenance. Who set that price? From which source? When? The cell says nothing.
• No real-time sync. Every price change has to be retyped into your shop and into every marketplace.
• No error monitoring. DIM_MM 620x820x450 and an empty category slip through just as easily as clean data.

The hidden costs that are not on your tab

Bad product data feels free, because the error is hidden. The bill arrives later, somewhere else. Research by Akeneo (The Evolution of the Modern Shopper, 2025, 1,800 consumers across eight countries) found that 40% of consumers returned an online purchase because of incorrect product information, and 53% abandoned a purchase because the data was wrong. Dissatisfaction with incomplete product data more than doubled: from 13% in 2023 to 30% in 2025.

And it does not stop at conversion. A slow product page costs you in Google too. Core Web Vitals, including Largest Contentful Paint, are a confirmed ranking signal (developers.google.com/search). Google calls an LCP up to 2.5 seconds good, and anything above 4 seconds poor. Those four uncompressed supplier photos of 4 MB each? You pay them back in load time, and therefore in ranking.

Why a PIM often becomes a six-month migration project

The standard reaction is: then just stand up a PIM. And on paper that is right. In practice it is rarely just. According to vendors like AtroPIM, a PIM implementation for a mid-sized catalogue typically takes three to six months. A project with 10,000 to 50,000 SKUs quickly runs to twelve to fourteen weeks from discovery to go-live.

The real snake in the grass is the data migration. It is consistently the most underestimated cost. Connecting a PIM to your ERP, webshop and DAM often costs three to five times the licence fees, thanks to middleware, custom mapping and maintenance. And poor adoption and unclear ownership sit behind a large share of failed enterprise software projects. You are not buying a tool, you are buying a project. And that project is exactly why many companies stay stuck in Excel.

From source to golden record

Here is how that triple office chair becomes one clean record. SyncRefine matches duplicate products on SKU, EAN and barcode, backed up by AI, and merges them into a golden record with per-field provenance, exactly the route that merging supplier data into one record describes step by step. Spellings are normalised: say 47 different colour labels are brought back to your twelve standard colours, so purple, lilac and violet all become a tidy Purple.

The result is a clean, sell-ready product catalogue that you do not have to reassemble for each channel. Add a source or change a price, and the rest follows.

Sell-ready also means fast, multilingual and monitored

A clean catalogue is more than tidy rows. Every supplier photo is converted to WebP at exactly the size the placement calls for, up to around 98% lighter files. No background removed, no watermark taken out, no AI upscaling: just lighter and faster, and that counts towards your Core Web Vitals.

AI enrichment writes missing descriptions, fills in attributes, categorises and translates into 42 languages. It learns your structure, not your tone of voice, so your brand voice stays yours. And anomalies, like an empty category or an EAN that does not check out, are automatically held back before they reach your shop. The live connection with WooCommerce, Magento and the marketplaces means you only have to get it right once.

You are not starting from zero

That is perhaps the most important part. You already have a shop, already have a structure, already have years of work in your range. A living catalogue does not mean throwing that away and starting over. It means your existing setup is the starting point, and that your sources connect neatly onto it. Take Giga Meubel as an example: a furniture wholesaler with 40 suppliers and 70,000 products, live in sync. No six-month project, but a catalogue that keeps itself up to date.