Knowledegable Master Data
Posted by Sastry at 3/5/2008 9:41 AM 
Categories: Business Web 2.0, Retail, Semantic Web, Master Data
Categories: Business Web 2.0, Retail, Semantic Web, Master Data
We have been discussing master data over the last few weeks and our belief that there are ways to improve the access to the specific data for a specific customer.
In the retail industry, we (the industry in standards efforts) made a very difficult improvement to the use of standardized codes and numbers. People not familiar with the specifics of retail data will still recognize the bar codes that are on products in supermarkets, mass merchants, category killers and department stores – basically everywhere. Getting order and integrity to the use of standardized codes has been a major success story in business history.
Other industries believe that they do not have the need for such specific standardized product identification. Although, last week when I was trying to find a particular type of Matt board to frame some of my photography, I wish that everyone did identify their products this well. It took quite a while to find “very white” board that matched what I was looking for – even though I know it was a specific model number from a specific manufacturer.
And where do I find an eSATA interface card for my Power Mac G5 that is a few years old – what is the specific card that works for the specific drive that I want to attach to my computer? The clerks at the stores that I have been too, have just looked at me and said – “What?”
Regardless of whether standardized codes are used for product identification, we believe that data quality – and customer satisfaction - can be improved if the knowledge about the use of the product is available and machine readable so that the products can be searched and the proper attributes delivered to the customer. Of course machine readable data is also searchable be the staff. I knew the card that I wanted had not shipped as yet; I was just having fun with the store staff of the retailer. The point is he did not have access to the knowledge. If he did, I would buy it form them when available – as it is I will buy it on-line.
Fortunately, there is significant progress being made in the storage and retrieval of knowledge about the products that are in the supply chain. This knowledge, properly stored and retrieved can eliminate much if not - all of the - quality problem that comes up with trying to use data from multiple systems fit it into customers requests for information.
In my Matt board example, the distributor that answered my text based inquiries first got the order – not a big deal it was about $300 – but the last time I looked getting a sale is better than losing a sale. For that industry – one that is “artistic” in nature - human (even text email messaging) interaction is essential. I needed to ask questions of the “knowledge plane” of those companies – the customer service department. In the world of data provisioning, staging and alignment – that request is automated and needs to be packages into to standardized or defined messages.
The technology for the storage of this knowledge about the data is in development and use is under discussion. We believe that it is time to start looking at how a knowledge plane above the relation and distributed data about master data can solve the problem of “dirty data” or incomplete data in the data synchronization efforts in the b2b supply chain.
We like to think of this “knowledge plane” as the semantic plane – and it sits on top of disparate data sources – like price information, ingredient or bill of materials, image, warranty, detailed specifications and the like.
The knowledge about the items is stored in that plane. When a price request by a customer is submitted the information for that customer is created based on knowledge about the programs in place, contracts for that customer, supply of the item, transportation surcharges and anything else specific to the supply chain in use.
When an image is requested the image for the particular version of the product that the customer needs and the use case for the image and the dimension data is fulfilled – so the plan-o-gram image (the small images used for in-store “maps” of the shelf in a Supermarket or Mass Merchant) – represents the product currently being shipped to the customer – not last years image and size data – or next years but the current one. Of course the customer should be able to request either. Also, if a high resolution EPS (Encapsulated Postscript File) is needed for an ad, it needs to be the one for the product that will be found in the supply chain when the ad is running.
Other examples, I am sure come to mind, but I believe that the point is clear. All of this data is in different systems – and usually in different locations. There are different security mechanisms in place for much of the data. But the knowledge of what is the authentic source of information for a particular request is probably distributed around the enterprise in the heads of associates. The semantic plane will store that knowledge and to the degree possible answer the request accurately. When not possible there are other Web 2.0 social networking tools that can alert the person that has to figure out the request and put together the data construct for the customer request. One of these tools – Twitter was discussed on Monday in our Blog.
We believe that concepts like the Semantic Plane and the use of social networking tools inside of the Enterprise can bring about a significant improvement in the quality and accuracy of the data being offered to customers. For more information about this approach please contact us at sales@dharacg.com.
Fred Geiger
www.dharacg.com
In the retail industry, we (the industry in standards efforts) made a very difficult improvement to the use of standardized codes and numbers. People not familiar with the specifics of retail data will still recognize the bar codes that are on products in supermarkets, mass merchants, category killers and department stores – basically everywhere. Getting order and integrity to the use of standardized codes has been a major success story in business history.
Other industries believe that they do not have the need for such specific standardized product identification. Although, last week when I was trying to find a particular type of Matt board to frame some of my photography, I wish that everyone did identify their products this well. It took quite a while to find “very white” board that matched what I was looking for – even though I know it was a specific model number from a specific manufacturer.
And where do I find an eSATA interface card for my Power Mac G5 that is a few years old – what is the specific card that works for the specific drive that I want to attach to my computer? The clerks at the stores that I have been too, have just looked at me and said – “What?”
Regardless of whether standardized codes are used for product identification, we believe that data quality – and customer satisfaction - can be improved if the knowledge about the use of the product is available and machine readable so that the products can be searched and the proper attributes delivered to the customer. Of course machine readable data is also searchable be the staff. I knew the card that I wanted had not shipped as yet; I was just having fun with the store staff of the retailer. The point is he did not have access to the knowledge. If he did, I would buy it form them when available – as it is I will buy it on-line.
Fortunately, there is significant progress being made in the storage and retrieval of knowledge about the products that are in the supply chain. This knowledge, properly stored and retrieved can eliminate much if not - all of the - quality problem that comes up with trying to use data from multiple systems fit it into customers requests for information.
In my Matt board example, the distributor that answered my text based inquiries first got the order – not a big deal it was about $300 – but the last time I looked getting a sale is better than losing a sale. For that industry – one that is “artistic” in nature - human (even text email messaging) interaction is essential. I needed to ask questions of the “knowledge plane” of those companies – the customer service department. In the world of data provisioning, staging and alignment – that request is automated and needs to be packages into to standardized or defined messages.
The technology for the storage of this knowledge about the data is in development and use is under discussion. We believe that it is time to start looking at how a knowledge plane above the relation and distributed data about master data can solve the problem of “dirty data” or incomplete data in the data synchronization efforts in the b2b supply chain.
We like to think of this “knowledge plane” as the semantic plane – and it sits on top of disparate data sources – like price information, ingredient or bill of materials, image, warranty, detailed specifications and the like.
The knowledge about the items is stored in that plane. When a price request by a customer is submitted the information for that customer is created based on knowledge about the programs in place, contracts for that customer, supply of the item, transportation surcharges and anything else specific to the supply chain in use.
When an image is requested the image for the particular version of the product that the customer needs and the use case for the image and the dimension data is fulfilled – so the plan-o-gram image (the small images used for in-store “maps” of the shelf in a Supermarket or Mass Merchant) – represents the product currently being shipped to the customer – not last years image and size data – or next years but the current one. Of course the customer should be able to request either. Also, if a high resolution EPS (Encapsulated Postscript File) is needed for an ad, it needs to be the one for the product that will be found in the supply chain when the ad is running.
Other examples, I am sure come to mind, but I believe that the point is clear. All of this data is in different systems – and usually in different locations. There are different security mechanisms in place for much of the data. But the knowledge of what is the authentic source of information for a particular request is probably distributed around the enterprise in the heads of associates. The semantic plane will store that knowledge and to the degree possible answer the request accurately. When not possible there are other Web 2.0 social networking tools that can alert the person that has to figure out the request and put together the data construct for the customer request. One of these tools – Twitter was discussed on Monday in our Blog.
We believe that concepts like the Semantic Plane and the use of social networking tools inside of the Enterprise can bring about a significant improvement in the quality and accuracy of the data being offered to customers. For more information about this approach please contact us at sales@dharacg.com.
Fred Geiger
www.dharacg.com
Comments