If an autoload file has been processed but a listing is missing or still shows an old price, start with one item. Save its identifier, file version and processing result. Then follow the path from source data to the listing page. This locates the failure without blindly changing the whole catalogue.
The result is a diagnostic log: what you expected, what the source actually contains, what reached the XML and what Avito reports. For the first check, you need access to your own catalogue and report, the URL of the file in use and one affected item. There is no need to change the schedule, republish the entire range or connect an API. Avito is a Russian-market classifieds platform; use the requirements for your actual category and account.
Separate four independent results
A feed is a file that transfers listing data. XML defines the structure of that data. Valid structure lets a program read the file, but does not confirm compliance with a particular category or publication rules.
| Result | Evidence | Next question |
|---|---|---|
| XML is readable | Syntax check | Does the data fit the category? |
| File processed | Result of the relevant upload | What happened to this item? |
| Listing published | Status and listing page | Are the terms correct? |
| Enquiry received | Enquiry record | What was the outcome? |
Record these results separately. “The export works” is too broad for acceptance: a file can be readable while carrying the price of a different configuration. Published listings do not prove enquiries or sales either. Those need their own records.
Save the initial state
Choose one affected item and one working control item in the same category. Record the internal SKU, the file's Id and the Avito listing number, if it already exists. These identifiers are different: one belongs to your system, one links the feed record and one identifies the platform listing.
Save the affected upload's time, complete message, file URL and generation date. Include the time zone. Copy the file before changing it so that you can compare the data that may actually have been used. Do not attach the entire file to a public discussion: it may contain operational information.
Record when you first noticed the issue and what changed beforehand: catalogue structure, export, category, images or file-storage domain. Timing helps choose a test, but does not itself prove causation.
Check the exact URL in the settings
Copy the address actually used by autoload and open it while signed out of your system. Check that it returns a file with the expected data, rather than a login page, server message or empty response. Do not replace the URL with one you assume is correct before recording the original result.
Find the chosen Id in the downloaded file and compare the price, title and images with the source. If the source record is updated but the file still contains old information, investigate feed generation. Possible areas include item selection, the generation schedule, a saved copy and file-server caching.
Successfully opening the file in your browser does not prove that Avito received it at the required time. If the report describes a retrieval problem, ask the technical provider to match the request time with your server's response. Check access restrictions and redirects while keeping service credentials private.
Separate XML syntax from category rules
Syntax checks concern the way elements are written: opening and closing tags, permitted characters and encoding. If a check gives a line and column, start there and inspect neighbouring data. One unclosed element can prevent the remainder of the file from being read.
For example, a literal ampersand in an XML text value must be written as & so that it is read as a text character. The XML specification describes special-character escaping and structure. Fix escaping in the generator where possible; otherwise the next export will reintroduce the original problematic character.
Do not apply indiscriminate character replacement across the entire file: double-escaping can damage values already encoded correctly. If software creates the files, ask the developer to use XML-aware checks rather than a set of text replacements. For manual preparation, retain the original copy and validate after each small change.
Then open the current category requirements in the documentation linked from Avito. Check required fields, allowed values and image specifications. Instructions for a neighbouring category may not apply. The official Avito for Business video about XML autoload covers file preparation; the video was published in 2024 and does not replace current field requirements.
Compare the product one field at a time
- Category. Confirm that the chosen category matches the listing subject. Compare required fields with its current template.
- Model. Compare the SKU and configuration. Similar names may hide a different size or variant.
- Price. Compare the source, XML and listing. Clarify the sales unit: one item, a set or another permitted unit.
- Availability. Check the source and sales terms. Record how the company handles out-of-stock items under its category's rules.
- Images. Open image URLs without authentication, then compare the photographs themselves with the model.
- Description. Check that it matches the actual product, including any prices and terms repeated in the text.
Record the exact discrepancy: “The source is a two-item set, but the description gives a single-item price.” “Incorrect data” forces the provider to repeat the diagnosis. If several sources supply one field, appoint an authoritative source and define the update rule.
Check that the Id remains stable between file versions. An existing item should not receive a random new identifier on each generation. Do not change the Id merely to retry a fix: first establish which relationship you would break. When moving to another system, prepare a mapping table and agree on how existing listings will be handled.
Read the message for the specific item
Find the chosen Id in the processing result. Check time and version: it is easy to mistake an old error for the outcome of a corrected file. Save the complete message with the field name and submitted value. A short “category error” note may conceal different causes for different products.
If the message describes an invalid value, compare it with the category's current allowed values. For an image error, check both file retrieval and the image itself. If placement terms are involved, valid XML will not solve the issue: understand the requirement rather than changing the product or description to bypass a check.
Do not launch repeated uploads in succession while it is unclear which version is being processed. Wait for an understandable result under the existing operating procedure. If you still need the platform's help, prepare one diagnostic record instead of a long message claiming that “the entire XML does not work”.
Illustrative example: the price correction lasts only one day
Source: the price is for a set. After nightly generation, the file contains the single-item price again. A manually corrected file displays the set correctly until the next update. Hypothesis: the export reads a different field. Check: compare mapping rules and two consecutive feed versions. Accept the fix after another normal generation, when the chosen Id is unchanged and the price and unit match the source.
All details in this example are illustrative. It shows why manually correcting a published value may leave the cause untouched. Verification lasts until the next agreed update, rather than ending with “all done” immediately after saving a setting.
Accept the correction
First confirm the change in the source or generator. Then check the new file version, its processing result and the specific listing page. Compare price, set contents, photographs and terms. Keep the previous message beside the new state.
Include the working control item: changing a shared template must not damage its data. If several categories' exports changed, include at least one agreed example for each affected rule. Repeat reconciliation after the next normal update to confirm that the correction persists.
A short developer brief could read: “For the specified Ids, identify the price source, correct the mapping, preserve identifiers, show the before-and-after files, and check the processing result and visible price. Do not include other categories without a separate list.”
When you do not need a new integration
If the problem is one incorrect attribute, start with its source and export rules. A new service will not automatically repair an incorrect product record. If only an outdated image URL fails, assess a storage fix before replacing the whole upload arrangement.
Custom development is justified when the existing process cannot maintain required data, preserve relationships or show processing results. Before buying, describe those limitations using specific items and agree on tests. This public catalogue and autoload case study can help discuss scope without promising sales.
Glossary
- Feed: a data file supplied to a platform.
- XML: a structured data format.
- Id: an item's identifier in the file; retain its relationship to the product.
- Generator: software that builds the file from sources.
- Field mapping: the rule determining where each value comes from.
If you cannot locate the failure, you can discuss an autoload diagnosis. Prepare an anonymised extract, Id, time, complete message and expected value first.
Apply this to your task
Locate a discrepancy for one item and prepare a testable brief without uncontrolled repeat uploads.
Single-item autoload diagnostic log: open the worksheet · Download the CSV to complete. Free access, no registration. Use it yourself or share it with your provider.