What is a feed and why does it matter
Table of Contents
- Feed as the single source of truth about products
- Difference between an input feed and an output feed
- How the feed affects ad quality
Feed as the single source of truth about products
A feed (source file) is a structured data file — in XML, CSV, or Google Sheets format — that contains complete information about all products in your store. Sembot fetches it directly from your e-commerce platform or a provided URL and uses it to build its entire knowledge of your product catalog.
Each row (record) in the feed describes one product and contains its attributes, i.e., properties: name, price, description, image URL, availability, category, EAN, brand, and many others. These are the data that Sembot stores, processes, and forwards to advertising channels.
Why is the feed "the source of truth"? Sembot does not manage your store — it is unaware of any price change, product withdrawal, or new promotion until it sees it in the feed. If something changed in the store but the feed has not yet been fetched, Sembot still operates on the old data. This is why feed freshness directly determines the freshness of your ads.
Example: a store reduces the price of shoes from $299 to $199 on Wednesday at 10:00 AM. Sembot will fetch the feed according to the configured schedule (e.g., at 12:00 PM UTC). Until then, ads will display the old price of $299.
Difference between an input feed and an output feed
In Sembot, a feed exists in two forms: input and output. These are two distinct stages of the product data flow.
Source file (input feed)
The source file is the feed that Sembot fetches from an external source — from your store, e-commerce platform, or any URL. It is the raw material, the starting point of the entire process.
In the Feed Settings tab you can see a list of all source files in the project. For each source file, Sembot displays:
| Column | Meaning |
|---|---|
| Name | Your own label to help identify the file |
| URL | The address from which Sembot fetches the file |
| Products | Number of products loaded from the last parse |
| Processed | Date and time of the last file fetch |
| Status | Active / Inactive |
| Supplementary | Whether this is a primary or supplementary feed |
Primary feed vs. supplementary feed
Sembot supports two subtypes of source files:
- Primary — the main file with the full product catalog. Every project should have at least one.
- Supplementary — an additional file that extends or overrides data from the primary feed. Used, for example, when some attributes (marketing descriptions, additional images) are stored in a separate file from the sales data.
It is possible to set separate parsing schedules for the primary feed and the supplementary feed — allowing independent update schedules.
Output file (output feed)
The output file is the feed that Sembot generates and exposes externally — to an advertising channel (e.g., Google Merchant Center, a price comparison site). It is the end result: processed, filtered, and formatted data ready for consumption by a specific platform.
The output file is automatically updated immediately after the source file is fetched and processed through product rules. There is no need to manually trigger generation — it happens in the background as part of every processing cycle.
The output file table displays:
| Column | Meaning |
|---|---|
| Name | Label to help identify the output file |
| Format | CSV or XML — according to channel requirements |
| Products | Number of products currently present in this file |
| Status | Active / Inactive |
| URL | The address from which the advertising channel fetches the file |
Data flow diagram
Store / e-commerce platform
↓ (fetching on schedule)
Source file (input feed)
↓ (processing through rules)
Sembot — rules and filters engine
↓ (automatic generation)
Output file (output feed)
↓ (fetching by channel)
Google Merchant Center / price comparison site / other channelOne project can have multiple source files (e.g., a separate feed for each language or market) and multiple output files (e.g., one for Google Shopping, one for a price comparison site, one for Facebook).
How the feed affects ad quality
The quality of product ads — their relevance, cost, and effectiveness — depends directly on the quality of the data in the feed. Sembot cannot improve an ad if the feed provides incomplete or outdated information.
Data completeness
Each advertising channel defines a set of attributes a product must have for an ad to be shown at all. Google Shopping requires, among others, a title, price, availability, product URL, and image URL. The absence of any of these fields causes the product to be rejected.
Sembot signals such issues in the Feed Settings section with alerts:
- This feed contains errors — products are being rejected by the channel
- This feed contains warnings — products are accepted but with incomplete data, which may reduce their visibility
- Missing required fields in the output feed — the mapping does not cover fields required by the channel
Data freshness
A product ad displays price and availability directly from the feed. If a product is "Out of Stock" in the store but the feed has not been fetched yet, the ad still directs customers to an unavailable product — leading to rejections by Google and a lower ad account rating.
The parsing frequency (set per project in the Feed Settings tab) determines how often Sembot fetches a new version of the source file. The more frequent the parsing, the faster store changes are reflected in ads.
Quality of descriptions and titles
A product title in the feed becomes directly the headline of a product ad in Google Shopping. Google's algorithm matches ads to user queries based on the title and description from the feed. Products with precise, rich titles (brand + model + size + color) achieve better query matching and higher CTR than products with generic descriptions.
Sembot allows these attributes to be modified through product rules before the output file is generated — without modifying the store's source data.
Product count vs. account limit
Sembot displays a counter in the Feed Settings tab: Active products in output files — showing how many products are currently active across all output files relative to the limit available under your plan. Exceeding or approaching the limit (a notification appears when 10% of the limit remains) may result in some products being excluded from ads.