Plan a controlled transfer of your Trendyol catalogue into an owned store, covering field mapping, variants, price, inventory, media and validation.
Moving products from Trendyol into an owned e-commerce channel is not a matter of copying titles and prices from one file to another. A dependable migration reconciles product identity, category attributes, variant relationships, inventory, tax, media rights and post-order operations under one data model. The objective is not to shut down marketplace sales; it is to open an additional channel in which the brand controls its catalogue.
Trendyol's official developer documentation presents V2 as the current integration area. Endpoints and field requirements can change, so implementation must always follow the current V2 documentation and the permissions assigned to the merchant account. Screen scraping, automated copying from the consumer interface or unauthorised data extraction are not sustainable integration methods.
First determine what can be transferred
Separate three classes of data:
- Brand master data: the brand's own SKU, barcode, product name, description, cost, tax rate and media for which it holds usage rights.
- Channel data: Trendyol product or merchant identifiers, category mapping, channel price and marketplace-specific status codes.
- Data that may belong to the platform or third parties: customer reviews, ratings, badges, page layout and platform-produced assets.
Build the new store around the first class. Retain the second only for mapping and reconciliation. Explicitly verify rights before moving anything in the third class. In particular, do not assume that ratings and customer reviews are catalogue content the brand may freely reproduce.
Create a source-to-target mapping dictionary
Before migration, prepare a data dictionary approved by all owners:
| Source concept | Target field | Transformation rule | Error behaviour |
|---|---|---|---|
| merchantSku | canonicalSku | Trim whitespace; preserve case | Quarantine when missing |
| barcode | barcode | Preserve as text; keep leading zeroes | Stop on duplication |
| productMainId | productGroupKey | Group variants under one product | Report orphan variants |
| categoryId | targetCategoryId | Use an approved mapping table | Keep as draft if unmatched |
| attributes | attributes | Normalise code, value and unit | Do not publish if required data is missing |
| salePrice | channelPrice | Apply tax and margin rules | Request approval below threshold |
| quantity | sourceInventory | Deduct safety stock; floor at zero | Raise anomaly alert |
| imageUrls | mediaAssets | Check HTTPS access, format and rights | Do not publish failed media |
Even a product title requires judgement. A keyword-heavy marketplace title may not suit the owned store's editorial voice. Keep a neutral name in the canonical catalogue and generate channel-specific titles. Apply the same principle to descriptions, SEO copy and delivery promises.
The variant model is the critical point
Colour, size and volume choices may appear as separate product rows in the source, while the target store presents them as choices on one product detail page. Preserve the individual SKU, barcode, stock, price and media relationship for every variant.
Reducing a “Black / 38” variant to “38”, for example, introduces a wrong-item fulfilment risk. A variant key can be a deterministic combination of colour and size codes, but it must remain stable. Test the parent product and variant matrix for these conditions:
- Do two variants in one group have the same option combination?
- Is one barcode attached to multiple active SKUs?
- Does every variant have sellable stock and a price?
- Does an image belong to the parent product or a particular colour?
- Is an inactive source variant accidentally reopened in the target?
A safer integration architecture
A direct “read from Trendyol, write to live store” pipeline is brittle. Add an observable staging layer:
Authorised source connection → raw data zone → transformation/mapping → validation → canonical catalogue → store publisher
A time-stamped raw response helps incident investigation, but unnecessary personal or secret data should not be retained and access must be restricted. Transformations must be repeatable: processing the same source version twice must not create duplicate products. A combination of source merchant identity, SKU and data version can serve as an idempotency key.
The connector must respect official limits, retry transient faults with increasing delay and avoid endlessly retrying permanent 4xx errors. Credentials do not belong in code or a feed URL; keep them in secret management and grant least privilege. The merchant account using the integration must be explicitly authorised to do so.
Separate initial migration from continuous synchronisation
The first catalogue transfer is a project; inventory and price updates are ongoing operations.
- Initial load: ingest product body, variants, attributes and media as drafts.
- Inventory sync: update sellable quantity at short intervals and apply a channel safety buffer.
- Price sync: designate a finance-approved price source; automated changes may not break the margin floor.
- Content sync: queue changes for review rather than automatically overwriting rich owned-store copy.
- Deactivation: do not erase an item when it disappears or closes at the source; make it inactive while retaining open-order and reporting links.
When both channels sell the same physical stock, “last known quantity” is insufficient. You need reservation between add-to-basket and payment, release after cancellation and an oversell policy. For critical products, separate channel allocation or a safety buffer is the simplest protection.
Three validation gates
Gate 1 — schema: required fields, type, length, allowed values, uniqueness and reference relationships.
Gate 2 — commercial: VAT, list/sale price, minimum contribution margin, inventory, sale eligibility and media rights.
Gate 3 — experience: mobile rendering, variant selection, add to basket, search/filter, delivery information and checkout.
The automated reconciliation report should at least show source product count versus ingested rows; unique SKU and barcode count; parent/variant totals; published, draft and quarantined products; inaccessible images; and price or inventory anomalies. Add human sampling by category.
Worked migration example
Consider a home-and-living brand with 600 Trendyol product rows. Discovery shows that these represent 420 parents and 180 variants. The first wave is limited to 80 parents with sales in the last 90 days and dependable inventory. Its category map defines 12 target categories, 46 required attributes and seven unit conversions. Products enter a draft store; 11 missing barcodes, eight duplicate SKUs and 17 inaccessible images go to quarantine. Nothing is published until these failures are resolved. Approval opens 10 products first and then all 80; 20 test orders verify inventory, shipping and returns end to end.
RACI: who owns each migration decision?
| Workstream | Responsible | Accountable | Consulted | Informed |
|---|---|---|---|---|
| Official connection and security | Integration/technology | Technology lead | Trendyol account owner | E-commerce |
| Field and category mapping | Catalogue operations | E-commerce lead | Product, SEO | Support |
| Price and tax | Finance | Finance lead | E-commerce | Management |
| Media and copy rights | Brand/content | Brand lead | Legal | Catalogue team |
| Inventory and orders | Operations | Operations lead | Technology, finance | Support |
| Go-live decision | Project manager | E-commerce lead | All workstream owners | Company |
Go-live checklist
- Have the current official V2 documentation and account permissions been confirmed?
- Are source identity and brand SKU separate and traceable?
- Are category and attribute maps versioned and approved?
- Have the variant matrix, barcode uniqueness and media relationship been validated?
- Have price, tax, margin floor and inventory safety rules been tested?
- Can the integration rerun without generating duplicate products?
- Is there a named owner for the error queue, alert, retry and manual correction?
- Has at least one cancellation, return and stock-out scenario been tested with a realistic order?
Once the catalogue migration is stable, do more than list products in the new store. Connect selected collections to creator storefronts to build demand for the new channel. Use the creator storefront feed guide for that next step.
Start e-commerce seller onboarding or create a brand account to bring your catalogue into an owned channel and influencer sales network through a controlled process.