Marketplace product listing at scale

Most integration tools on the market function as a simple "switch." If you sell goods with a known EAN code that already exists on a marketplace, the system merely attaches your offer to an existing product card. However, the real challenge begins where standard catalogs end.

This case study demonstrates how Orchesty can be used to master one of the most demanding disciplines in e-commerce integration: the automated creation of a complete product identity from scratch.

When This Strategy Becomes Essential #

This advanced orchestration model is designed for scenarios where conventional automation fails:

In-house Production and Private Labels: You sell unique products under your own brand that do not yet exist in marketplace catalogs such as Amazon, Allegro, or Kaufland.
Unlisted EAN Codes: The marketplace does not recognize your product, and you must be the one to create it, including all categories, descriptions, and technical parameters.
Specific or Dynamic Assortment: Handcrafted goods, art, custom manufacturing, or fashion, where every season brings thousands of new items that cannot be matched to existing templates.

Architecture: Orchesty as the "Master of Attributes" #

In this scenario, Orchesty does not just move data. It acts as the orchestrator for a central repository of marketplace data. The ERP remains the system of record for inventory and pricing, while Orchesty manages the specific technical parameters required by various sales channels.

Marketplace product listing flow covering category selection, attribute enrichment, validation, and Store persistence — Orchesty orchestrates category assignment, attribute completion, validation, and multi-channel storage in a single listing flow.

Phase A: Launch and Routing #

The process can be initiated in two ways: in real time via Webhook when a product is created, or in bulk by mapping the entire catalog. The first critical checkpoint is the verification: Does the product already have an assigned category?

If NO (New Product):
- Vector Search: Orchesty converts the product name into a vector and performs a lightning-fast search to find the most semantically similar categories among thousands of marketplace options.
- AI Selection: An AI model such as Gemini makes the final selection of the Category ID from a narrow list of candidates.
- Store Category: The result is stored for future use.
If YES (Update): The topology bypasses this phase, saving both time and AI processing costs. This path is crucial for situations where the marketplace changes the requirements for mandatory attributes within already listed categories.

Phase B: Intelligent Extraction and Decision Tables #

Once the category is known, Orchesty identifies the current marketplace requirements for attributes while using caching for maximum performance.

AI Set Attributes: Orchesty uses AI to analyze product descriptions from the ERP. The AI extracts specific values such as material, closure type, or season from unstructured text.
Decision Tables (Google Sheets): For data not present in the description, such as brand mapping, warranty periods, or country of origin, Orchesty uses decision tables managed in Google Sheets. These tables are cached within Orchesty, allowing rapid completion of static parameters without manual mapper configuration.

Phase C: Multi-channel Store and Finalization #

This is the most vital part of the process, ensuring data cleanliness throughout the ecosystem.

Final Validation: Before the product is sent, Orchesty checks whether all mandatory attributes are completed. If the product fails validation, the process stops and the administrator is notified.
Create Marketplace Product: Upon successful validation, the product is created directly via the marketplace API.
Store Marketplace ID and Attributes: Orchesty saves the Marketplace Product ID and all specific marketplace attributes into a dedicated Store service.

Why Store Attributes Outside the ERP

ERP Integrity: The ERP is not burdened with thousands of technical parameters required only by the marketplace, such as specific Allegro or Amazon classifications.
Multi-channel Readiness: You build a database within the Store that can be shared across channels. If you later expand to another marketplace, Orchesty already has the product parameters ready and only needs to re-map them to the new schema.
Speed: Accessing data in a standalone Store is significantly faster than repeated queries to the ERP or marketplace APIs.

Conclusion: A Resilient and Scalable Gateway #

Thanks to this topology, listing becomes an automated assembly line. The system is intelligent enough to handle new products through AI and Vector Search while remaining robust enough to efficiently update existing ones. A dedicated Attribute Store then becomes a strategic asset for the company, enabling rapid expansion into additional global markets.

ID mapping guide — the Marketplace Product ID stored in the Attribute Store is exactly the cross-system mapping pattern this guide describes.
Data comparator guide — for "Phase A bypass" updates, the comparator skips products whose marketplace-relevant attributes did not actually change.
API caching guide — the same primitive used to cache decision tables (Google Sheets) and category requirements per marketplace.
Marketplace order processing — sister use case for what happens once the listed products start receiving orders.