How to Match Product Catalogs Across Amazon, Shopify, and eBay (When SKUs Don't Match)

If you sell on more than one marketplace, you already know the problem. You have a product catalog on Shopify. You have a listings feed on Amazon. You have an eBay store. Each one says it has the same products. None of them line up.

The SKUs are different. The titles are different. The brand names are formatted five different ways. Even when a product has a UPC, half your listings don't have it filled in.

So when you try to answer a basic question — "am I selling this same product across all three channels?" — the answer is "I have no idea, someone would have to open a spreadsheet for a week."

This post walks through a much faster way. Export your catalog from each marketplace, run fuzzy matching on title plus brand plus identifier, and get a single reconciled view in about 15 minutes. Here's how.

Why SKUs Don't Match Across Marketplaces

The core issue: every marketplace has its own identifier system, and none of them talks to the others.

In theory you can set your Shopify SKU to be the same string as your Amazon seller SKU and your eBay custom label. In practice, sellers set SKUs by hand years ago, changed their SKU scheme twice, and now half the listings have BLU-MED-2019, some have BLUE-M, and one has bluemedium-v2.

UPCs and GTINs are supposed to be the universal identifier. They work great when they're filled in. But most catalogs have UPCs on 60-80% of listings, missing on the rest — especially private label, handmade, custom, or bundled products. Matching on UPC alone leaves a big chunk of your catalog unreconciled.

The Real Cost of an Unreconciled Catalog

Sellers underestimate this because the cost is spread across a dozen small problems. Add them up and it's usually the largest hidden expense in a multi-channel operation.

• Overselling. You have 12 units in your warehouse. Amazon thinks it has 10, Shopify thinks it has 8, eBay thinks it has 5. You get orders on all three, and now you're apologising to a buyer.
• Wrong price on the wrong channel. You raise the price on Shopify but forget the Amazon listing is a different SKU, so you keep selling it there for a loss.
• Duplicate marketing spend. You run ads for the same product on two channels, each targeting different SKUs, competing with yourself.
• Poor forecasting. Your inventory report says you sold 40 units this month, but really you sold 65 across all channels — the reporting tool only saw one channel's SKU.
• Onboarding a new marketplace takes weeks. Every time you add Walmart or TikTok Shop, you re-do the whole SKU cleanup from scratch.

None of this is a headline problem. All of it is quietly eating margin.

The Approach: Multi-Column Fuzzy Matching

Instead of trying to match on a single field, match on three at once:

1. Product title as the primary column, weighted 55-65%.
2. Brand or manufacturer as the first helper, weighted 20-30%.
3. UPC / GTIN / MPN as the second helper, weighted 10-20%.

The idea is simple. Titles are noisy but always present. Brand is usually clean and disambiguates products with similar names. UPC is a strong signal when it exists and gets ignored gracefully when it doesn't. Blending all three gets you much higher accuracy than any one of them alone.

Step 1: Export Each Marketplace to CSV

You need one CSV per marketplace. Here's the fastest way to get them:

• Amazon: Seller Central → Inventory → Manage All Inventory → Download the "Active Listings Report" (or the Category Listings Report if you need more fields). Includes ASIN, seller SKU, title, brand, and (if you filled it) UPC.
• Shopify: Products → Export → All products, CSV for Excel. Includes SKU, title, vendor (brand), and barcode.
• eBay: Seller Hub → Listings → Download → Active. Includes item ID, custom label (SKU), title, brand, and UPC where filled in.

Rename the columns before you match so they're consistent across files. At minimum, you want title, brand, upc, and sku in each file. Don't worry about anything else yet — you can enrich later.

Step 2: Pick Your "Source of Truth" File

Two-file matching needs a left file and a right file. Pick your cleanest catalog as the left (the "source of truth") and match every other file against it, one at a time.

Most sellers pick Shopify as source of truth because it's the one they control most directly. If your Amazon listings are your primary revenue source and your Shopify is more of an afterthought, flip it. It doesn't really matter as long as you're consistent.

Step 3: Run the Match — Shopify vs Amazon

Upload both files into a fuzzy matching tool. Configure it like this:

Hit Match. You'll get a list of pairs with a similarity score. Anything above 85 is almost always the same product. Anything below 65 is almost never the same product. The 65-85 zone is your review pile — usually 5-15% of your catalog, which is small enough to check by eye.

Step 4: Add eBay (or Any Other Channel)

Repeat step 3 with your source-of-truth file (Shopify) on the left and eBay on the right. Same configuration, same weights. Merge the two output files by Shopify SKU and you have a single reconciled view:

This one table is what an MDM (master data management) platform charges $30,000/year to give you. You just built it in a fuzzy matching tool in 15 minutes.

Handling Product Variants

Variants (size, color, material) are where naive matching breaks. "T-Shirt" on Amazon and "T-Shirt" on Shopify both match at 100% on title — but if one is a size Medium and the other is size XL, they're different SKUs.

The fix: add size and color as low-weight helper columns.

• Primary: title — 50%
• Helper 1: brand — 20%
• Helper 2: size — 15%
• Helper 3: color — 15%

Now "Blue Widget Pro Medium" ↔ "Blue Widget Pro Medium" scores near 100. "Blue Widget Pro Medium" ↔ "Blue Widget Pro Large" gets penalised by the size mismatch and drops below the match threshold. The tool learns to treat variants as separate SKUs without you writing any code.

What to Do With the Reconciled Catalog

Once you have the merged table, you unlock a bunch of things that were previously painful:

1. Sync inventory correctly. Push a single stock number to all three channels. If you have an inventory tool, feed it the reconciled SKU map.
2. Enforce consistent pricing. Set one price per product and use the map to update every channel at once.
3. Find gaps. Products on Shopify but not on Amazon are expansion opportunities. Products on eBay only are probably legacy listings you forgot about.
4. Find duplicates within a channel. Sometimes the same product is listed twice on Amazon with slightly different titles. Multi-column matching finds those too.
5. Onboard new channels faster. Next time you add Walmart, you already have a canonical master list to map into.

Common Mistakes to Avoid

Trying to match SKUs directly. If your SKUs aren't identical across channels, direct matching will find almost nothing. Match on title plus brand plus UPC instead — SKU is one of the least reliable fields.

Using UPC alone. UPCs are missing on a big chunk of most catalogs. Match on UPC as a helper column, not as the primary. It'll push scores up when present and gracefully do nothing when absent.

Setting the match threshold too low. If you accept everything above 50%, you'll bulk-merge products that aren't the same. Start at 85%, then walk down through the 70-85 band manually.

Doing all channels at once. Match pairwise (Shopify vs Amazon, then Shopify vs eBay). Trying to match three files in one pass is confusing and hard to review. Pairwise is cleaner.

Forgetting variants. If you sell in multiple sizes or colors, add those as helper columns from day one. Otherwise you'll merge XL and Small into "one" product and break your inventory sync.

How Often to Rerun

Once a month is usually enough for stable catalogs. If you add new listings frequently, run it weekly. The whole pipeline (export, upload, match, review, merge) takes under an hour once you've done it twice — the setup work is one-time, and future runs are basically re-uploading fresh CSVs and clicking match.

Keep the reconciled master list somewhere your team can see. A Google Sheet works. An Airtable base works better. If you have an MDM tool or PIM, use it — but you don't need one to start.

The Bigger Picture

Multi-channel e-commerce sellers spend an enormous amount of time fighting their own catalogs. Not because the work is hard, but because the tools are exact-match and the data is fuzzy. Every marketplace names things slightly differently, and the mismatch compounds every year.

Fuzzy matching on title plus brand plus identifier collapses that whole problem into one afternoon of work per month. You end up with a single reconciled view that makes inventory, pricing, and forecasting decisions much easier — and you get most of the benefit of an enterprise MDM system for the cost of a spreadsheet tool.

Amazon, Shopify, and eBay each want you to think of them as the centre of your business. They're not. The centre is your product catalog, and that catalog only exists once you can match it across all three.