Choose the catalogue ImgAff can use

Prepare a product feed for useful image matches

Your feed defines the products and affiliate links available to ImgAff. This guide shows what the file must contain and how supported exports are read.

The four things every product needs

Your file is a spreadsheet — CSV or another delimited text format — with one row per product and a header row naming the columns. Four of those columns are required:

What	Why
A product ID	A stable code identifying the product. We use it to tell an updated product from a new one when you re-upload.
A product name	What the product is called. This is most of what we match against your images.
An affiliate link	Your tracked link — the one that earns you commission. Must start with `http://` or `https://`.
An image URL	A link to the product's photo.

Every product needs an image URL This is the rule that catches people out. ImgAff works by matching pictures, so a product with no photo can never be shown on your page — those rows are skipped. If your export has no image column, you need a different export; the products cannot be used without one.

Use the tracked link, not the shop's own page Some exports contain both: a link to the product on the merchant's website, and your tracked affiliate link. They look almost identical. If you upload the merchant's own link, everything will appear to work — products will show on your pages and visitors will click through and buy — but you will earn nothing, and nothing will warn you. When in doubt, the tracked link is usually the longer one, and usually points at the network's domain rather than the shop's.

Useful, but optional

Include these if your export has them. Nothing breaks if it doesn't.

Price and currency — shown on the product overlay. A price we can't read is simply left off; the product still works.
Description — helps matching accuracy.
Category, brand, SKU — stored, and useful later.

A minimal example

If you're building the file yourself, this is genuinely all you need:

id,name,link,image_url,price,currency
1001,Blue Ceramic Mug,https://example-network.com/click?p=1001,https://shop.example/img/1001.jpg,12.50,EUR
1002,Walnut Serving Board,https://example-network.com/click?p=1002,https://shop.example/img/1002.jpg,29.00,EUR

You don't have to use these exact column names — see below.

Where to get your file

Pick your network when you upload, and we read its usual column names automatically. You do not need to rename anything.

Awin

Download your product feed as a delimited file (comma, tab or pipe all work). We read aw_product_id, product_name, aw_deep_link and merchant_image_url, along with search_price, currency and merchant_name. Note that aw_deep_link is your tracked link — merchant_deep_link is not.

Rakuten Advertising

Download the delimited product catalog. Rakuten's publisher catalog has no header row — it starts with a line beginning HDR| and ends with one beginning TRL| — which is normal and we handle it. Just upload the file as it comes.

impact.com

Download the catalog as CSV or TAB. We read catalog item id, product name, product url and image url, in either the spaced or the underscored spelling. Both the advertiser-format and impact-format downloads work.

CJ (Commission Junction)

Download the product catalog as comma- or tab-delimited. We read id, name, buyurl and imageurl.

Your own shop (Shopify, WooCommerce, anything else)

Choose Generic CSV when you upload. We recognise the common names — id, title, link, image_url, plus price, currency, description, brand and category.

Two things to know about a Shopify export Shopify writes one row per image, so a product with five photos becomes five rows, with the name blank on all but the first. Only the first row of each product will import — the others are reported as skipped, which is expected and not a problem.

More importantly, a Shopify product export contains no affiliate link column — it is an export of your own catalogue, not of tracked links. You will need to add a column containing the link you want clicks to go to.

Amazon Associates

Amazon does not give affiliates a product feed to download. There is no export button to find, and this is the single most common thing people go looking for. Your options are:

Build a small CSV by hand — perfectly reasonable for the handful of products you actually feature in an article. Use the example above; take the image URL from the product page.
Use a tool that already holds your Amazon products — several WordPress plugins collect Amazon products into your own site, and can export a CSV that our Generic CSV option reads.
Amazon's API — access requires an approved Associates account with qualifying sales, and Amazon has been changing which API it offers. If you already have API access, this is the route; if you don't, one of the options above will be faster.

We read text files, not XML

Several networks offer both a delimited text file and an XML version of the same catalog. Choose the delimited one — CSV, TSV or pipe-delimited. We do not read XML.

Large files are fine: up to 25 MB, and you can gzip them (.csv.gz). Feeds can also be given to us as a URL, in which case we re-fetch them automatically.

If your file is rejected

Two different things can happen, and the difference matters.

The whole file is refused

This means a required column was missing, or the file didn't match the network you picked. The message names the columns we actually found, which is usually enough to spot the problem — most often it's the wrong network selected. Your existing products are left untouched.

Some rows are skipped

The rest of the file still imports. Common reasons:

No image URL — the product can't be matched to a photo.
The link isn't a proper web address — it must start with http:// or https://.
Duplicate product ID — the later row is used.

The upload summary lists each reason with a count and the line numbers, so you can fix the file where the problem actually is.

Ready to connect the catalogue?

Open the publisher portal to upload the feed. If an export is not covered here, send a small sample to [email protected] and we will confirm what the current importer can read.

Open publisher portal