Smart Fields

What are Smart Fields?

Every data source uses its own field names. One source may call a product title title, while another uses name or product_name. Smart Fields map these different source fields to one fixed internal product structure.

Examples include title, description, brand, SKU, EAN/barcode, price, stock, image, and product type. Once mapped, Nifelo understands each value and can use the standardized data for imports, categorization, AI enrichment, normalization, validation, rules, and exports.

This guide explains the general operation of Smart Fields. The same central guide can be opened from several places in the data-source menu, keeping the fundamentals in one location.

The exact mapping may differ by source type. After reading this guide, also open Smart Fields under the source type you use. If a dedicated guide is available, it contains the additional or different workflow for that source. Follow the source-specific guide where the instructions differ.

Automatic or manual mapping

For file, feed, and non-managed webshop sources, names and structures may differ by supplier. Nifelo detects the available fields, after which you verify which Smart Field each source field should populate.

Use Auto-Map when available to create an initial proposal, but always review it before saving.

A technically valid mapping is not automatically correct. Make sure, for example, that a sales price is not mapped to a cost price and that an EAN does not contain an internal product ID.

Status and progress

The status shows mapped fields, missing required Smart Fields, unused source fields, and overall progress. Not every technical field needs a mapping; skip fields that have no purpose in products, AI, validation, or exports.

Reading a mapping

Each mapping has a target Smart Field and a source field. Example: product_name → title.

Use the sample product and raw source data to inspect the actual values. Review several products because one sample may be empty or exceptional.

Ways to build a value

Map

Use one source field directly, for example vendor → brand.

Combine

Combine two source fields when information is consistently split across them.

Fallback

Use a second field when the first is empty, for example sale_price with price as fallback.

Static value

Assign the same value to every product, such as condition new when this is true for the entire source.

Only use a static value when it genuinely applies to every product.

Add or change a mapping

1. Select Add mapping.
2. Choose the target Smart Field.
3. Select map, combine, fallback, or static value.
4. Choose the source field or enter the fixed value.
5. Review the data flow or live preview.
6. Check several products.
7. Save only when the results are correct.

A Smart Field normally has one active primary mapping per source. Check whether the target already exists before adding another.

Unmapped source fields

Map a field when it contains important product information, is needed for AI or validation, is required by a sales channel, or will be used in a rule or export. Leave technical metadata and irrelevant fields unmapped.

If an important value does not fit the standard library, an additional field may be created when that feature is available. Use a clear, stable name and avoid duplicate fields with the same meaning.

Using rules

Rules modify values under specific conditions: if a condition matches, then perform an action. Example: if the title contains NIKE, set brand to Nike.

Keep rules specific and test them with multiple products. Broad rules can change more products than intended.

Keep the number of rules limited

Do not repair product data with an ever-growing collection of rules. Many rules are difficult to understand, may conflict, and require maintenance when the assortment changes.

Prefer this order:

1. Correct the source data when information is structurally wrong or incomplete.
2. Use AI for content decisions where meaning, context, or varying wording matters.
3. Use a rule for a targeted, predictable exception that must always produce the same result.

AI is suited to improving titles and descriptions or interpreting product information. Rules are better for exact agreements, such as mapping a fixed supplier code to the same brand.

Rules remain a powerful way to intervene when the source cannot be corrected, AI lacks sufficient confidence, or an outcome must be strict and repeatable. Keep every rule small, explainable, and testable.

Save and re-import

Select Save mappings after reviewing them. Saving does not modify the original webshop, feed, or file; it controls how Nifelo processes the data internally.

Changed mappings may only become fully visible after the source is processed or imported again.

Reset mappings

Only reset mappings when you deliberately want to start over. Manual mappings may be lost. Record required fields and custom rules before major changes.

Recommended first check

1. Check title, product and variant IDs, SKU, EAN/barcode, brand, price, and image.
2. Resolve missing required fields.
3. Review Auto-Map proposals.
4. Inspect several sample products and raw data.
5. Use fallback and static values only when necessary.
6. Test rules on different products.
7. Save, import again, and inspect the products before using AI, normalization, or exports.