Opti-X - Prepare your Data Ecommerce Inventory

💡Note that these fields need to map accordingly with the real time events . For more information [click here]

Summary

OptiX data is designed to support three hierarchical levels, in order: product, variant and sku.

Clients may not have variant or sku level information, therefore see explanation to fill this out.

Use the RAW_DATA field to supply additional data which does not fit the schema. These fields are used to filter and output additional fields in the placement responses.

The 'Explanation' and Opti-X Impact' columns describes the implications of missing data.

FieldsMandatoryTypeExampleExplanationOpti-X Impact
SKU_IDNoStringRippedJeans1934256-Blue-W28H30Unique ID representing the lowest level of a product.

If not applicable, set this to value of VARIANT_ID.
Used in recommendation logic, clients may want to configure showing items at a SKU level.
VARIANT_IDNoStringRippedJeans1934256-BlueUnique ID representing the variant of a product.

If not applicable, set this to value of PRODUCT_ID.
Used in recommendation logic, clients may want to configure showing items at a Variant level.

Placement Recommendation response output.
PRODUCT_IDYesStringRippedJeans1934256Unique ID representing the product.Foundational to base personalised recommendations with.

Placement Recommendation response output.
ITEM_NAMEYesStringRipped JeansName of the product.Placement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
ITEM_DESCRIPTIONNoString100% Soft Denim Fabric!Description of the itemPlacement Recommendation response output.
SKU_NAMENoStringBlue Ripped Jeans, W28 H30Name of the SKUPlacement Recommendation response output.
VARIANT_NAMENoStringBlue Ripped JeansName of the variantPlacement Recommendation response output.
CATEGORYIDYesString050102Unique ID representing the category of the product.Placement Recommendation response output.
CATEGORYYesStringJeansName of the category.Placement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
DEPARTMENTIDYesString05Unique ID representing the department.Placement Recommendation response output.
DEPARTMENTYesStringClothingUnique ID representing the department name, highest level of categoryPlacement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
CATEGORY_LISTYesStringClothing > Trousers > JeansOrder of category name where the highest is the department and lowest is the category. Please use '>' delimiter.Placement Recommndation Ecomm Rules logic to identify categories of parent categories.
CATEGORY_KEY_LISTYesString05 > 0501 > 050102Order of category key where the highest is the department and lowest is the category. Please use '>' delimiter.Placement Recommndation Ecomm Rules logic to identify categories of parent categories.
BRANDYesStringExample BrandBrandPlacement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
CURRENCYNo, if only one currency.StringUSDCurrencyCurrency of prices.
PRICENo (Recommended, see impact)Number40Standard price before saleE-commerce Rules (Filtering recommendations within a placement).

Price related smart-campaign use-cases
SIZENoStringW28 H30Size. Not always applicable.Placement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
COLOURNoStringBlueColour. Not always applicable.Placement Recommendation response output.

E-commerce Rules (Filtering recommendations within a placement)
STOCKNo (Recommended, see impact)Number826Current StockE-commerce Rules (Filtering recommendations within a placement).

Stock related use-cases.
RELEASE_DATENo (Recommended, see impact)String2024-10-15When the product was released/shown on the e-commerce site.Required for Model/Methods and smart campaign use cases related to the release date, e.g. showing new items.
LAST_UPDATEDNoString2024-10-16 08:02:30When this item was last updated.Sanity checking internal mapping and tracking changes.
FEATURESNoStringJeans, Ripped, Blue, BaggyFeatures of the item.Placement Recommendation response output.

It may be relevant for training new models/methods.
DISPLAY_STATUSNoBooleanTRUEWhether the item is displayed. Default set to TRUE if customer is not providing data.E-commerce Rules (Filtering recommendations within a placement)
DISCONTINUE_DATENoString2024-10-20Date of discontinued item.Sanity checking internal mapping and tracking changes.
ISDISCONTINUEDNoBooleanFALSEWhether the item is discontinued. Default set to FALSE if customer is not providing data.E-commerce Rules (Filtering recommendations within a placement)
ISAVAILABLENoBooleanTRUEHard Boolean representing if the item should still be displayed in recommendations. Default set to TRUE if customer is not providing data.If the item in inventory is always expected to return in recommendations.
RAW_DATANoList of dictionaries/jsonIn the format of:

[
{'attribute_key': Value1,
'attribute_value': Value2,
'level': product|variant|sku,
'type': text|numeric|boolean}
]

Example:
[{'attribute_key': is_sale, 'attribute_value': TRUE, 'level': sku, 'type':boolean}, {'attribute_key': type, 'attribute_value': store_only, 'level': product, 'type':text}]
List of JSON containing additional data.

Value1 e.g. is_mini
Value2 e.g. 1
level e.g. either product|variant|sku
type: text|numeric|boolean
Extra data used to prefilter and output in a Placement Recommendation Response.
PRODUCT_URLNo (Recommended, see impact)Stringhttps://www.example.com/products/afjoi239Useful for Image Based ModelsResponse field of recommend, smart search and smart campaigns.
PRODUCT_IMAGENo (Recommended, see impact)Stringhttps://www.example.com/products/afjoi239.pngE-commerce Rules (Filtering recommendations within a placement)

Recommended for Smart Banner use-cases
Response field of recommend, smart search and smart campaigns.
VIDEO_URLNoStringhttps://www.example.com/products/afjoi239.movCan be useful in other use-cases. Possibly smart banner addition.
RATINGNoNumber3Numeric value, no range specified.Determined by the client.