Synthetic Industry

Troubleshooting guide · updated 2026-10-11

Merchant Center cannot process your product file? Check the format, header row and XML layout

The file-level rules that make Merchant Center reject or misread a product file, from accepted formats, header row, tabs and line breaks, encoding, XML namespace and tag case, extensions and compression.

File-level errors come before item-level errors

Merchant Center processes a file in two stages. First it has to read the file. Only then does it check each product against its rules about prices, availability and images. A fault in the first stage can hide everything in the second, which is why a data source with file-level errors is the first thing to fix, and why fixing item problems while the file is unreadable accomplishes nothing. Google's help pages change; this guide reflects the pages as read on 11 October 2026.

  • If the latest-update report lists file-level errors, stop looking at item diagnostics.
  • If items are read but flagged, the cause is in the data, not the file layout.

Text files: header row, tabs and line breaks

Google accepts tab-delimited text with a .txt or .tsv extension. The first row holds attribute names, one per column, and each product is a row. An empty cell is fine when an attribute has no value, but an extra or missing tab shifts everything after it, which the troubleshooting page names as the usual cause of failure. A tab or a line break inside a description does the same, so a field that may contain them must be cleaned before export. The expected encodings are UTF-8 and UTF-16, and other encodings can cause processing errors, which is the same byte problem described in the guide on character encoding. Excel's own file types are not accepted; export to tab-delimited text first.

XML files: namespace, tag case and escaping

The XML examples use RSS 2.0, with Atom 1.0 also named. The root declares the Google namespace, http://base.google.com/ns/1.0, a channel holds the title, link and description, and each product is an item whose attributes are g:-prefixed tags such as g:id and g:price. Attribute names must be lowercase, so title is accepted and Title is not, and a literal ampersand in a product name must be escaped. Those two points are ordinary XML rules: end-tag names must match exactly and ampersands must be escaped, as the guide on well-formed XML explains. Save the file with an .xml extension; a file with no extension is read as tab-delimited.

Extension, compression and size

The extension tells Google how to read the file. A compressed file must really be compressed, holding exactly one .txt or .xml file; a file named .zip that is not a zip archive fails. The page lists 4 GB per file, and the troubleshooting page lists a lower figure of 500 MB for compressed files, so check the current limits for your case. The address must lead directly to the file, because an address that leads to a web page makes processing fail.

Required attributes and fixed values

Some attributes accept only fixed values. Availability takes in_stock, out_of_stock, preorder or backorder, and condition takes new, refurbished or used; writing brand new for condition is a listed mistake. Google's troubleshooting page calls the id attribute one of the most commonly forgotten, so check it first. Google's file page lists availability as in_stock, out_of_stock, preorder or backorder, yet its own XML sample writes the value as in stock with a space, so follow the list of allowed values rather than copying a sample. A spreadsheet-based export that silently converts a value can cause these faults too.

A safe first investigation

Open a copy of the file in a plain text editor. For a text file, look at the first line for attribute names separated by tabs and count the tabs on a few lines; they should match. For XML, look for the rss or feed root with the Google namespace declared and for unescaped ampersands. Compare the processed item count in the data source report with the number of rows or items in the file. Send redacted screenshots of the report, never a login.

How the paid job is accepted

The job merchant-center-feed-file-fetch-processing-errors is £245 for one data source and one hosted file. The address must return the file itself with no intermediate page, with no login for a public file or with the login your account holder has stored in Merchant Center settings; after the next fetch the latest-update report must show no file-level errors; and the number of items read must equal the rows or items in the file, with every difference listed. Your account holder applies changes and triggers processing, and we never hold your Merchant Center login. Prices are untested proposals, and payment follows the agreed checks and your sign-off. Nothing is booked or charged by an enquiry. Google's review and recrawl timing and decisions are its own, and approval is never promised.

Sources and limits

  • Google Merchant Center Help: create a product file Checked 2026-10-11.
    • Tab-delimited text (.txt or .tsv) and XML are accepted; unsupported types such as .html prevent processing; a file link must start with http://, https:// or sftp://; one file can be up to 4 GB.
    • A text file puts attribute names in the first row, one product per row; stray tabs or line breaks inside a field cause formatting errors; UTF-8 and UTF-16 are the expected encodings.
    • The XML examples use RSS 2.0 with the g namespace http://base.google.com/ns/1.0, a channel, and an item per product with g:-prefixed attributes; Atom 1.0 is also named.
    • Omitting fundamental attributes such as id, title, price, availability or link is listed among the common feed-processing errors; condition must be new, refurbished or used and availability must be in_stock, out_of_stock, preorder or backorder, and entering brand new instead of new is the page's example of an invalid value.
    • The XML sample on the same page writes <g:availability>in stock</g:availability> with a space, while the list of allowed values and the tab-delimited examples use in_stock.
  • Google Merchant Center Help: troubleshoot my data source Checked 2026-10-11.
    • Excel files are not accepted; RSS and Atom files should end in .xml and a file with no extension is read as tab-delimited; compressed files must really be compressed and hold exactly one file; the id attribute is called one of the most commonly forgotten attributes.
    • In XML, special characters such as & cannot appear unescaped and attribute names must be lowercase; an extra or missing tab is the usual tab-delimited fault.
    • Direct upload, SFTP and scheduled fetch files must each be under 4 GB; compressed files under 500 MB.