Page tree
Skip to end of metadata
Go to start of metadata

As mentioned on the page Prepare the Product Data, this part of the documentation describes the requirements that your product data feed needs to fulfill in order to build Product Guides entirely yourself with the excentos Workbench.

If you have more complex feeds (e.g. web services or XML feed), please return to the Prepare the Product Data section or contact excentos support if you have any questions regarding your data - we'll help you cover any data feed requirements.


Table of Contents:

Upload Requirements and Limitations


  • The file has to be a CSV (comma-separated values) file. Only this file type is accepted.
  • If your product data is written in an Excel sheet (.xls or .xlsx) or in an OpenDocument format (.ods) which are used by LibreOffice or OpenOffice, the file has to be exported to a CSV file (see Export the Product Data File into CSV).

Data Feed File Requirements


If you want to connect your product data directly to the Workbench, the product data CSV file needs to fulfill the following requirements (as mentioned above, please contact excentos for more complex formats such as XML or JSON):

Type / FieldSpecification
File typeCSV file (preferred encoded is UTF-8)
Field separatorNeeds to be one of the following: ',' (comma), ';' (semi-colon), ':' (colon), '|' (pipe), tab or whitespace character.
Text delimiterText cells have to be surrounded by " " (double quote) or ' ' (single quote). One of them at the beginning of a cell; one at the end of a cell. This is especially important if the text contains a specified field separator.
Escape characterThe escape character that is used within the text cells. Normally this is a '\' (backslash), to escape a single double quote, another single double quote can also be used.
Decimal separatorDefault is a '.' (dot), e.g. '799.99'. For comma separated decimals (e.g. '799,99') you need to set the number format in the Workbench.
Multiple values separator

If some attributes contain cells with more than one value in a cell (e.g. the attribute color with the values black and blue), these values need to be separated by a separator (e.g. with a comma such as "black,blue"). Otherwise these two values will be interpreted as one value and cannot be split by using the "Multi Value" switch button in the data attribute table (see Upload your Product Data File, Multi Values).

You can change the global multi values separator anytime as well as the attribute-specific separator in the Product Data view.

Units

Such as kilogram, currencies etc. should be in a separate field if the data field is required to be interpreted as a number.

(we've marked this as suggestion only because the unit is only required if it shall be displayed in the Product Guide. The unit is not technically required, except if prices are displayed in the result list because we then need to know the currency)


Hint

The field separator and text delimiter as well as the file type encoding and the escape character can be set during the update process of your product data file.


Required Fields and Their Names


The Workbench expects a fixed name convention for some core product attributes.

  • The "Required" attributes in the table below are the minimal set of data that every Product Guide needs to work properly.
  • The "Recommended" attributes are technically optional, but we recommend you using them in order to display certain information like the product name or price in the product recommendation card.


Attribute Mapping

Despite that fixed naming convention it is fortunately possible to map the underlying attribute names to your product attribute naming schema (see Attribute Mapping).

Hint

(lightbulb) If you build up your product data from scratch, the easiest way to fulfill the data requirements is to start with the Template Data File.


General Data Attributes

(info) General product attributes are identical across all product categories of your Shop / Website.


The required and recommended data attributes and the name conventions for the data fields, which can come from several data sources are:

Attribute NameRequiredRecommendedRemarksVisible in Recommendation
id

X

-

Unique ID for a single product

If your feed does not contain that kind of information, an internal generic id is created to distinguish all products. This internal id is not visible in the data or the Workbench at all.

-
name

(X)

X

Name of the product

The name is not technically required, but makes it a lot easier for you and the end-users to identify products in the result list.

(tick)
image

(X)

X

Product image URL, e.g. a URL to the product image on your website

The image is not technically required, but makes it a lot easier for you and the end-users to identify products in the result list.

(tick)
price

 -

XProduct price (decimal number)(tick)
manufacturer

 -

XManufacturer or brand name(tick)
detailUrl

 -

XURL to the product details page(tick)
purchaseUrl

 -

(X)

URL to shopping basket. We recommend including it since the product recommendation can then directly perform a "add to basket" action without the user needing to visit the product details page first. However, this depends on the desired customer journey throughout your shop / website.

This feature has to be supported by your website and furthermore implemented for your individual requirements.

(tick)
variantId

 -

-Master variant ID of a product. Useful if your product assortment contains e.g. color or size variants. excentos can collect them to make sure only the master product is displayed in the product result list.-
description-XProduct description text(tick)
currency unit
(X)Not strictly necessary, but cleaner. The currency can be directly included in the price attribute, too.
status / available / online etc
(X)If you can not restrict your product feed to products actually available, please provide a suitable filter attribute to excentos
category
(X)Having a category indicator for the product is often helpful. Please try to include the tree of parent categories (in a separate attribute if possible)
Push factors / merchandising information
(X)If you have information available that helps excentos to optimize the product listing for your business (such as margin, stock levels), include them.

All attribute names are case sensitive

That means that the name 'price' and 'Price' will be recognized as two different attributes (with small and capital 'p'). Please make sure that all attributes are written correctly to match the required/recommended attributes above.

Otherwise make sure that your own attribute names are mapped correctly. Else the corresponding information cannot be displayed or executed properly.

(lightbulb) For more information on possible data attributes and which data might help to even improve the Product Guide, please also check https://www.excentos.com/en/documentation/data-feed-requirements.

Product Data Sources

This part of the documentation describes different Product Data sources. The Product Data can consist of several data sources and can also be connected by different connector types (HTTP, FTP, SFTP).

excentos can help to preprocess these Product Data sources to use them for your Product Guide. You can contact excentos support if you have any questions regarding your Product Data sources.

Possible Product Data sources

SourceDescriptionRecommendedOptional

PIM 

product information like part numbers, variant id, descriptions etc.x
technical attributescharacteristics of a product; can be contained in master PIM or in another databasex
availability feeddefines product availability per region
x
country-specific product definitionse.g. different certificates per product; can be contained in master PIM or in other database
x

Category-specific Product Data

Category-Specific Attributes describe the product itself, such as technical data, possible use-cases and everything that distinguishes one product from the other. For every Product Guide, we need all the category specific attributes like e.g. the screen resolution of a notebook computer, the shoe size of a mountaineering boot, the range of height adjustment of an office desk etc. The more data we have, the more powerful and detailed the Product Guide can be.

The category-specific attributes can have the same data structure (mostly the case with CSV files) or a separate data structure / different attributes per category (mostly the case with XML data structures).

Category-Specific Attributes are not required for the data sourcing option „third-party licensed data“.

Example of category-specific attributes

Typical CSV file with product data for, e.g., bicycles:

product data example csv raw


This is how the CSV file needs to look like if you open it in e.g. Excel if you have correctly set the text delimiters (in the example above we used a comma):

product data example csv table

Data Feed for international rollouts


You can either provide

  • individual feeds per market / country / etc.
  • or tell excentos a filter criterion within a central feed from which excentos can decide in which market to include the product.

If prices, product names and descriptions differ, we recommend providing separate data feeds for each market (although it would be technically possible to contain all markets in one feed).


Mime Type Detection


During the upload process the Workbench is analyzing the data file type. Only if the file is a valid CSV the file can be uploaded. The corresponding mime type (internal format) for CSV files is 'text/csv'. Unfortunately, there are 600+ words, word phrases or hexadecimal values (mostly at the start of the file) that can affect the mime type. That means that a word like 'Article' at the beginning of a CSV file results in the mime type 'message/news'. With that resulting type, it is not possible to upload the CSV anymore although the file is completely valid. Because of the huge amount of words and word phrases, excentos is not able to prevent these errors during the upload process. Nevertheless, the mayor part of these words/phrases are not relevant for a CSV file use case, like detecting python or image files.

If you have any problems regarding this issue do not hesitate to contact us. Some "bad words" examples are listed below:

Bad wordsMime Type
Article
message/news
Path:
message/news
[version]
application/inf
StartFontMetrics
application/x-font-adobe-metric
Rar!
application/x-rar-compressed
Spreadsheet
application/x-sc
FWS
application/x-shockwave-flash
CWS
application/x-shockwave-flash
PSID
audio/prs.sid
BEGMF
image/cgm
icns
image/icns
AGD2
image/x-freehand
FreeHand10
image/x-freehand
MATLAB
application/x-matlab-data
#!/bin/perl
text/x-perl
#!/bin/python
text/x-python
FLV
video/x-flv
MOVI
video/x-sgi-movie
\037\213
application/gzip

Mime Type XML

Read about the Template Data File and Example Data File to get started building your own Data File

  • No labels