Product Feed

Provide general information about your products. We require (but not limit to) only a small set of data. Here is described its structure and format.

You can either

  • compile product data into a single CSV file and give it to us
  • publish a regularly updated file under a specific secret link and give is the link.
    In the latter case Sizolution robot will regularly download the product data and synchronize our databases. This method allows us to make frequent automatic updates (up to several times a day). For security reasons do not to publish the link to your feed. Make it as hard as possible to find or guess.

We prefer - but do not require - JSON format. Other options are XML and CSV. If you have a small online shop with less than 1000 articles, CSV probably is your best choice.

If you already have a product feed for another purpose that contains the following information, we can probably work with it as well - our developers have to look at it first.


Examples: feed.json, feed.xml, feed.csv.
List of countries
List of clothing materials

Product feed can be represented as a list of objects (rows), one for each product_id. Required fields (columns) are marked with *, the rest are optional, but very helpful for getting the best performance of Sizolution.

Product Identity

We identify a garment using two fields - product_id and size. Both are required to uniquely identify a particular piece of clothing that the user receives. In the product feed we ask only for meta information and not sizing, so here you provide only product_id. Sizing is provided in a separate feed.

product_id* Unique product identifier in your shop, independent of size. E.g. let it be "TH1234-42".
group_id Multiple variants or colors of the same product can be grouped together and have the same group_id but different variant, e.g. a "T-shirt with a unicorn" can be sold in "red", "green" and "black/white dotted" variants. Products can also be available in a single variant. In case you group products by color/variant, we'd like to have an identifier of the group.
link to the product page.

Model Identity

We link your data with our intelligence on brands. Provide the following information.
brand* E.g "Tommy Hilfiger".
article* Article number - brand's equivalent for product_id. E.g. "ww0ww23088251". Pair (brand, article) must uniquely identify the clothing.
model Sometimes called "style" - brand's equivalent for group_id. It groups different colors/variants. E.g. "ww0ww23088".
GTIN Global Trade Item Number or EAN or UPC, e.g. "0798512420477".

General Information

name Product name or a short description, e.g. "T-Shirt with a unicorn".
classifier* Ordered list of categories assigned to the clothing.
gender* Intended clothing gender - one of female, male or unisex.
age_group One of adult, child or newborn.
description E.g. "Men's exclusive T-Shirt with a colorful unicorn, made from a a slick cotton fabrick".
photos List of links to photos of the product, preferably at least 300x300 px.
country_designed_in A country from the list.
country_made_in A country from the list.

Material Information

material_type One of woven, knitted, denim, broadcloth or leather.
material_composition Dictionary {key: value} with keys form the list and values showing percentage, e.g. {'cotton': 99, 'elastane': 1}.
material_elasticity One of no, low, medium or high. High elasticity can be found on thermal clothing, leggins an so on.

Construction Details

lining: 0/1
elastic_bust_band: 0/1
elastic_waist_band: 0/1
elastic_belt_band: 0/1
elastic_hip_band: 0/1
fit One of tight, slim, regular, relaxed, loose or baggy.
season_spring: 0/1
season_summer: 0/1
season_autumn: 0/1
season_winter: 0/1

Sizing and Measurements

Provide a separate feed with sizing information and measurements.


Clothing Standard Sizing Scheme (CS3)
Examples: measurements.json, measurements.xml, measurements.csv.

Measurements feed can be represented as a list of objects (rows), one for each size for each product_id. Required fields (columns) are marked with *.


Product identifier, as stated in Product Feed.
product_id* E.g. "TH1234-42".


size_id Size identifier in your store, e.g. "M-176-1".
size* Original size provided by the brand, e.g. "96A-176".

If you incorporate some or multiple local sizing systems or you know how brand size is converted to other sizing systems, please provide this information as well, in form of additional fields (columns):
int_size: M
eu_size: 48


product_weight Net weight of the clothing.

If you don't use Sizolution Automatic Clothing Scanner:
cs3* Clothing measurements in CS3 format.
is_from_blueprint*: 0/1. Is cs3 taken from a blueprint of the garment (1), or measured by hand (0).

If you do use Sizolution Automatic Clothing Scanner, instead you should provide the following:
available*: 0/1 Whether this size is available in the storage. This parameter is needed to anticipate which sizes are going to be scanned.
barcodes* List of your inner barcodes, associated with this combination of size_id and product_id.

Of course you can combine these cases and have both sets of columns (fields) in your measurements feed.

Sales and Returns

For better fit predictions and size recommendations based on personal preferences of customers we also use sales and returns history, including history for items that were not measured.


Examples: sales.csv.

Sales history can be represented as a list of objects (rows). Required fields (columns) are marked with *.

General Order Information

order_id* Order identifier.
order_timestamp* Order date and time in UTC: "%Y-%m-%d %H:%M".
user_id* Same user id as on the website.
product_id* Product id as in Product Feed.

Model Identity

brand E.g "Tommy Hilfiger".
model E.g. "ww0ww23088".
article Article number.
classifier_text* Type of clothing (e.g. “Women-Dresses”).

Shipped size

size_id Size identifier in your store, as stated in [Sizing and Measurements][feed.html#sizing-and-measurements].
size* Original product size string.
eu_size If you incorporate some or multiple local sizing systems, provide the converted values on the moment of check out.
barcode of the ordered item.


currency_code* 3-letter IBAN currency code, e.g. "EUR".
checkout_price* Unit price after all discounts.
original_price* Unit price before any discounts.
promotion_name Name of promotion campaign for discount, e.g. "Black Friday".
n_items* Number of items of the same product_id and size in the order.

Order Status

status* Order status: processing, delivered, cancelled or returned.
user_comment If any. Reason for returning the item, if known.
user_satisfaction A score that the user gave to the item, 0 (disappointed) to 100 (satisfied).

Once a full sales and returns history is transfered it can be updated for the last 3 months once per day and be available on some unique link (similar to the feed).

A hack to enable ToC