Why PIM and eCommerce?
Data is power. Information is power. An important aspect of any eCommerce solution is to present as much data and information as possible to the shoppers. Shoppers love to visit websites which contain useful information for their exploration, research, studies or whatever it takes to help them decide on a purchase. Product Information Management (PIM) systems play a crucial role in managing the enterprise data. Integration of an eCommerce system with PIM is part of overall eCommerce strategy. Ensure a proper PIM solution is in place before the start of any eCommerce solution. Single view of products enables a better multi-channel experience by providing consistent, complete, and accurate product information to customers. It also provides complete and consistent product information across all commerce channels e.g. eCommerce, store, kiosk, mobile, print catalog, Global Data Synchronization (GDSN). This consideration is all the more important when an enterprise has a wide array of products, frequently changing product characteristics and non-uniform IT infrastructure. eCommerce often needs data in different flavors depending on the channel of sale. For example, mobile app may not require detailed product information but a desktop browser might require as much information as possible.
What PIM contains?
Typically PIM holds product information such as
- product name
- short to long product descriptions
- attribute and its values
- images of different sizes and its meta data
- catalog taxonomy
- workflows to represent business processes
IBM Commerce and PIM
IBM WebSphere Commerce has a strong integration story with Advanced Catalog Management (ACM) which is part of InfoSphere MDM (Master Data Management) suite. Advanced Catalog Management provides easy-to-use tools for business users to centrally manage a cross-channel strategy. Business users can create and manage precision marketing campaigns, promotions, catalog, and merchandising across all sales channels. WebSphere Commerce consumes the output from ACM to present a single, unified platform which offers the ability to do business directly with consumers, directly with businesses, and indirectly through channel partners (indirect business models).
ACM has a 3-step data synchronization.
1. Generating the Collaboration Server data from the InfoSphere MDM Collaboration Server
2. Mapping the InfoSphere MDM Collaboration Server data to the WebSphere Commerce data, and
3. Uploading the WebSphere Commerce data to the WebSphere Commerce Server.
Utilize the pre-configured data model in InfoSphere MDM Collaboration Server which mirrors the data model of WebSphere Commerce, it has some sample workflows (like New Product Introduction, Manage Catalog Entry), predefined user roles with appropriate access privileges as well as integration with WebSphere Commerce to export the catalog entries. These features can be extended or modified as per an organizations requirement.
There are 2 integration approaches with WebSphere Commerce server
- Web service approach
- Batch load approach
For a small volume of data, ACM uses real time a web service call to publish data directly to IBM WebSphere Commerce system. This approach is used for catalog groups and attribute dictionary exports.
For a large volume of data, ACM starts an asynchronous batch load process. This approach is used for catalog entries. The batch process uses WebSphere MQ for notifications and an FTP server for large data file transfer.
For non-IBM PIM solutions, the synchronization concepts with WebSphere Commerce described above remains pretty much the same. Utilize the sample dataload files given in WebSphere Commerce information center to model the output from the PIM system and use the OOB dataload tool to load the WebSphere Commerce business objects. Other option is to use BODL (Business Object Data Load), a services asset from IBM, to load the WC business objects. Some implementations involve customizing BODL to transform and enrich the data from PIM. In some of the recent implementations, both OOB dataload tool and BODL were used to achieve the complex dataload requirements.
To conclude, integrating with a PIM should be one of the key solution architecture pieces for any eCommerce engagements. Dataload must be treated as a “mini-project” and must be taken up for implementation in the very beginning of the overall eCommerce journey!