Setting Up a Supplier Adapter

An adapter is FeedPrep's way of remembering how to normalize data from a specific supplier. Once you set up an adapter, every future feed from that supplier is processed automatically using the same rules.

What Is an Adapter?

An adapter is a saved set of normalization rules tied to a specific supplier. It stores everything FeedPrep needs to transform that supplier's raw data into your canonical format:

• Column mappings — key-value pairs that translate the supplier's column names to your attribute names (e.g., colour_name → color)
• Value mappings — per-attribute mappings that translate supplier values to your approved values (e.g., Midnight → black for the color attribute)
• Unit conversions — rules for converting measurements to your preferred units
• Separator configuration — defines how multiselect values are delimited for this supplier (e.g., pipe |, semicolon ;, or comma)

Think of an adapter as a translation profile. It captures all the supplier-specific knowledge so you don't have to redo the work each time.

How to Create an Adapter

You don't create an adapter from scratch in a settings page. Instead, you build it naturally as you process a feed:

1. Upload and process a feed — Import a file from the supplier and open it for processing.
2. Map columns — For each of the supplier's columns, select the corresponding attribute from your schema. FeedPrep saves these column mappings to the adapter automatically.
3. Map values — For select and multiselect attributes, map the supplier's raw values to your approved values. These value mappings are saved per attribute on the adapter.

Every mapping you make is persisted to the adapter in real time. There's no separate "save" step — the adapter builds itself as you work.

Automatic Application

Once a supplier has an active adapter, any new feed uploaded from that supplier is automatically processed using the adapter's rules. The column mappings apply, the value mappings apply, and the normalization pipeline runs — all without manual intervention.

This is the "map it once, reuse forever" pattern. The first feed from a new supplier requires manual setup. Every subsequent feed from that supplier flows through automatically.

What the Adapter Stores

Here's a closer look at each component of an adapter:

Column Mappings

Stored as simple key-value pairs where the key is the supplier's raw column name and the value is your canonical attribute name. For example:

Value Mappings

Stored per attribute. Each attribute can have its own set of raw-to-normalized value mappings. For example, the color attribute might have:

Separator Configuration

For multiselect attributes, the adapter stores which separator the supplier uses to delimit multiple values within a single cell. Common separators include pipe (|), semicolon (;), and comma (,). FeedPrep can auto-detect the separator, but you can also set it explicitly on the adapter.

Editing an Adapter

You can view and modify a supplier's adapter at any time from the supplier detail page. Add new mappings, update existing ones, or remove mappings that are no longer relevant. Changes take effect the next time a feed from that supplier is processed.

What's Next?

With your adapter in place, FeedPrep runs your data through a 13-step normalization pipeline. Learn exactly what happens at each step in Understanding the Normalization Pipeline.