Skip to content

Product feed errors: how to catch catalog drops before campaigns lose products

A beginner-friendly routine for checking imports, item drops, warnings and practical alert thresholds.

J

Julia Villegas

· 9 min read · 255 views
Product path showing how to detect product feed errors before catalog drops

Product feed errors can remove products from your campaigns before anyone notices them in Meta, Google Shopping, TikTok or Pinterest. If you run an online store, checking imports, item drops and feed warnings helps you catch problems early instead of assuming the catalog is fine just because a file or URL was generated.

What product feed errors are

A catalog is the structured list of products you send to an advertising or commerce platform. It usually includes a product ID, title, description, price, availability, product image, product URL, brand, category and other attributes required by each channel.

A product feed is the file or URL that carries that catalog to a platform. It may be XML, CSV, TSV, JSON or another supported format. The basic idea is simple: your store holds the product data; the feed prepares that data so a channel can read it.

Product feed errors happen when that chain breaks or degrades. They are not always obvious at first. Sometimes the channel can still read the feed, but you lose product coverage, data quality or the ability to segment campaigns properly. For day-to-day operations, it helps to separate three signals:

Signal What it means Why it matters
Error A product or feed does not meet a required condition The item may not be usable in campaigns
Warning A data point is incomplete, weak or potentially problematic It may affect catalog quality or future reviews
Item drop The number of imported or exported products falls unexpectedly It may point to broken fields, excessive filters or incomplete imports

A typical error could be a product without a valid image. A warning could be a thin title or missing recommended category. An item drop may happen when a supplier changes a column name, when a rule filters more products than expected, or when an import arrives incomplete.

The goal is not only to check whether a feed exists. You also need to know what came in, what was transformed, what was excluded and which warnings keep appearing.

Why this matters for catalog ads

Catalog ads depend on product data being present, readable and available for the right channel. When the catalog fails, it is not just a technical issue. It affects marketing coverage, media budget, merchandising and the daily confidence of the team running campaigns.

You should pay close attention if you:

  • Manage hundreds, thousands or tens of thousands of products.
  • Change prices, offers or stock frequently.
  • Use the same source catalog for several advertising channels.
  • Activate and deactivate categories by season.
  • Import data from suppliers, marketplaces, ERP systems or external files.
  • Work at an agency and need to monitor multiple client catalogs quickly.

Example: imagine an apparel store starts the week with a healthy catalog. On Tuesday, a supplier changes the availability value from in_stock to available. The feed URL still works, and there is no obvious outage. But the rule that builds the campaign set no longer recognizes those products as in stock, so a large part of the catalog may disappear from the final feed.

That is the kind of problem you want to catch before it becomes a loss of campaign coverage. You cannot guarantee zero errors, because each channel applies its own checks and the frequency of your feed generation is not always the same as the frequency with which the channel reads it. What you can build is a practical control routine that reduces surprises.

Product feed errors: the signals to review first

To detect product feed errors in a practical way, start with five basic checks: import, mapping, item count variation, rules and channel warnings.

1. Import: check that the expected catalog arrived

The first question is not “was a feed generated?” The better question is “did the right catalog come in?” If you connect a store or product file, review whether the source was read correctly and whether key fields contain data.

Minimum fields to check:

  • Product ID or unique identifier.
  • Title.
  • Price.
  • Availability or stock.
  • Product URL.
  • Image URL.
  • Category or product type.
  • Brand, where relevant.

If one of these fields is empty across many products, the issue may be in the store, the source file, the field mapping or a later transformation.

2. Mapping: translate your fields into the channel’s language

Mapping means deciding which source field corresponds to each attribute required by a destination. Your store may call a field product_name, while a channel expects that value to become the product title. Your internal stock field may need to become an availability value accepted by the destination.

Mapping often works until someone changes the source file. If a supplier renames price as sale_price, your feed may lose the expected price unless you update the mapping.

A simple review is to open a representative sample of products and confirm that each attribute lands in the right place. You do not need to inspect every row manually. Start with products on promotion, out-of-stock products, products with variants and items from different categories.

3. Item drops: set thresholds before there is a crisis

An item drop happens when the final catalog contains fewer products than expected. Sometimes that is normal. You may intentionally exclude out-of-stock products, products without images or items outside a specific campaign category. The problem starts when the drop was not expected.

For beginners, a threshold is an internal alert rule: “if the number falls beyond what we consider normal, we review it.” There is no universal percentage that works for every business. A store with fast-moving inventory needs different tolerances from a retailer with a stable catalog.

Example: a furniture store usually imports between 12,000 and 12,500 products from its source. After excluding discontinued products and items without images, its Meta catalog feed normally contains between 10,500 and 11,000 items. If one day the final feed contains 6,200 products, it is not enough to say “the feed was generated.” The team needs to find out where the drop happened.

A simple control table can help:

Checkpoint Expected value What to review if it changes
Products in source Usual range from store or file Incomplete export, supplier issue, source connection
Products after cleanup Range after removing duplicates or invalid data Rules too broad, empty required fields
Products by set Normal range for each campaign or category Category filters, stock logic, price or brand rules
Products in final feed Acceptable range for each channel Mapping, format, destination requirements

Using ranges keeps you from overreacting to normal variation while still catching unusual changes.

4. Rules and sets: watch what you exclude

A rule transforms, completes or filters product data. For example: “exclude products without an image” or “mark products with a zero price for review.” A set is a subset of the catalog, such as a category, brand, price range or seasonal campaign group.

Rules make catalog operations manageable, but they can also cause unexpected drops if they are too broad. Before publishing important changes, review which rule changed, which products it affects, how many products remain before and after, and whether the final set still makes sense for the campaign.

This matters even more when one catalog feeds several channels. Meta, Google Shopping, TikTok, Pinterest or other destinations may need different formats, attributes or product sets. Applying exactly the same logic everywhere can create avoidable problems.

5. Warnings: do not wait for rejection

A warning does not always block a product, but it tells you something deserves attention. In large feeds, repeated warnings are usually more useful than isolated cases. If 20 products have broken image URLs, it may be a limited incident. If thousands share the same warning, there is probably a pattern.

Sort warnings by priority:

  • High: warnings that affect core data such as price, availability, image or URL.
  • Medium: warnings related to categorization, data quality or recommended attributes.
  • Low: smaller improvements that do not block the current operation.

Separating critical issues from planned improvements helps you act calmly instead of turning every feed review into an emergency.

How to review a catalog with AdMotion

AdMotion can help you build this review routine around the catalog rather than around manual spreadsheets. It connects product sources, maps fields, applies rules, creates sets and generates feed URLs for channels. The aim is not to guarantee that every platform will approve every product, because each platform runs its own review. The aim is to gain more control before you register or update the feed URL in the channel.

A practical workflow looks like this:

  1. Connect the catalog source. Start from a compatible store or a product file. The important point is to use a clear, repeatable source.
  2. Map the fields. Connect source fields to the attributes you need: title, price, availability, image, product link, category and other relevant data.
  3. Normalize with rules. Clean inconsistent values, complete data where appropriate and define sensible exclusions. Rules can also be written in natural language and reviewed before applying them.
  4. Create sets. Split the catalog into useful subsets, such as in-stock products, seasonal categories, gifts by price range or products for a specific campaign.
  5. Generate the feed by channel. Each destination may need its own format and attributes. With Export Channels, you can prepare feeds for Meta, Google Shopping, TikTok, Pinterest, Snapchat, Criteo, Microsoft Advertising, Thribee, ChatGPT or a custom channel.
  6. Review errors and warnings. The analyzer checks each feed against the channel specification. This does not mean guaranteed approval by the destination, but it helps surface issues before the feed is used.
  7. Copy the feed URL into the channel. The feed URL is registered in the relevant platform. Remember that the frequency of feed generation or download can be different from how often the channel reads it.

If you are just getting started, focus first on the Feed Manager: connect the catalog, map fields, apply rules and check that product counts do not collapse without explanation. The Feed Manager documentation page explains this role in the catalog workflow. Once the feed is stable, you can work on product-level visuals with the Templates Editor, especially if you want creatives that use catalog fields such as product photo, title, price, brand or availability.

AdMotion’s Export Channels page is also useful context if you send catalogs to several destinations, because each channel can have its own feed format, specification checks and product sets.

Example: a pre-campaign catalog review

Example: a sporting goods online store is preparing an autumn campaign. It sells running, hiking and fitness products. The team wants to activate catalog ads and send feeds to several channels.

During the review, they see this:

  • The source catalog contains 18,400 products.
  • After excluding products without images, 17,900 remain.
  • After filtering out-of-stock products, 13,700 remain.
  • The running set contains 4,200 products.
  • The final feed for one channel contains only 1,100 products.

The difference between 4,200 and 1,100 cannot be explained by stock or image exclusions. The team checks the rules and finds a condition that requires the exact category value “running.” However, the source catalog uses values such as “running shoes,” “running apparel” and “running accessories.” The rule was not careless, but it was too strict.

The answer is not to add products manually. It is to correct the set logic, review a sample of affected products, check the new count and analyze the feed again. After that, the team can copy the updated URL into the relevant channel.

This example shows why intermediate counts matter. If you only look at the final number, you know something is wrong. If you look at each step, you know where to investigate.

Final check before you trust a feed

Before a feed becomes part of your campaign routine, run a short review that covers both data and process.

Use this checklist:

  1. Source imported correctly. The expected catalog was read from the store or file.
  2. Core fields are populated. IDs, titles, prices, availability, URLs and images are present in a product sample.
  3. Mapping still makes sense. Source fields still match the destination attributes after any supplier, store or file change.
  4. Rules are explainable. Each exclusion has a clear reason and does not remove more products than expected.
  5. Sets have plausible counts. Campaign subsets match the range you expect for the season, stock level and category.
  6. Warnings are prioritized. Critical warnings are handled first; minor improvements are planned separately.
  7. Channel timing is understood. Your feed may update on one schedule while the destination reads it on another.

This does not create a zero-failure system. It creates a predictable review habit, which is what most teams need when catalogs become too large to inspect manually.

What to do this week

  1. Write down your normal counts. Record how many products you have in the source, after cleanup, by set and in the final feed. You do not need a complex system to begin; you need a baseline.
  2. Choose alert thresholds. Decide which drops deserve a review. Do this by catalog and by set, not only with one global number.
  3. Check critical fields. Review a sample that includes discounted products, variants, out-of-stock items and several categories. Confirm that title, image, URL, price and availability look right.
  4. Rank warnings by impact. Start with warnings that affect core data. Move cosmetic or lower-risk improvements to a planned review.
  5. Document rule changes. Every time you change a rule or set, write down what you expected to happen and compare it with the real count.

If you want to centralize this routine, review AdMotion’s Feed Manager. It lets you connect a catalog, clean it with rules, split it into sets and publish feed URLs for channels. You can also try it with your own catalog if you need to see where product drops happen before scaling the process.

Catching product feed errors is not about chasing problems after campaigns have already lost coverage. It is about building a simple, repeatable review: what enters, what changes, what is excluded and what finally reaches each channel.

Frequently asked questions

What is the difference between a product feed error and a warning?

An error usually means a product or feed fails a required condition, such as a missing image, invalid URL or empty mandatory field. A warning points to something incomplete or potentially risky, but it may not block the item immediately. Review both, but prioritize issues affecting price, availability, image and product URL.

Why can the number of products in a feed suddenly drop?

A sudden drop can come from real stock changes, renamed source fields, incomplete imports, strict rules or sets that no longer match the catalog. Compare several checkpoints: products in the source, products after cleanup, products by set and products in the final feed. That helps you find where the loss happens.

Does a feed analyzer guarantee that every channel will approve my products?

No. A feed analyzer can help detect errors and warnings against a channel specification before you use the feed, but it does not replace the platform’s own review. Also, your feed generation schedule and the channel’s reading schedule can differ, so instant synchronization should not be assumed.

Sources

  1. Feed Manager de AdMotion — AdMotion
  2. Export Channels de AdMotion — AdMotion
Share:
J

Written by

Julia Villegas

Performance Marketing

Julia Villegas works in performance marketing at AdMotion.

Related posts