Your SlideShare is downloading. ×
Better Retailing through Linked Data
Upcoming SlideShare
Loading in...5
×

Thanks for flagging this SlideShare!

Oops! An error has occurred.

×

Introducing the official SlideShare app

Stunning, full-screen experience for iPhone and Android

Text the download link to your phone

Standard text messaging rates apply

Better Retailing through Linked Data

1,581
views

Published on

Opportunities, perspectives, and vision on Linked Data in retail …

Opportunities, perspectives, and vision on Linked Data in retail

Published in: Technology, Business

1 Comment
5 Likes
Statistics
Notes
  • Annotations/ transcript:
    Slide 2: Products can be complex objects, and the product landscape is vast -- millions if not billions of products available to the average consumer.  Products in the Best Buy catalogue have an average of 31 actionable attributes (attributes that can influence a purchasing decision) ...overall actionable attributes range from a couple, in the case of a simple object like a gift card, to near 100 in the case of a home appliance. While consumers have a literal paradox of choice, a very small percentage of products make up a majority of consumer purchases and company revenue (estimated between 10-15%), leaving 85-90% of product offerings undiscovered or underutilized. This discrepancy is something I’ve called the “product long tail”, where a minority of retail products make up a majority of customer purchases and company revenue.

    Slide 3: One of the reasons I believe there is such a long tail of under sold products is that we haven't used the set of built in complex relationships that products naturally have with one another to our advantage. Products can be linked together in a number of different ways using linked data to form relationships. These relationships can be explored for the benefit of the consumer and the company. Its a level of product discovery that goes way beyond 'people who purchased this item also bought ” scenario that is so common on many retail web sites.  

    Slide 4: With these complex product offerings and deep product relationships, humans (both consumers and employees) are at a disadvantage. As a consumer have you ever been challenged to find that one product that exactly fits your needs? Or have you experienced an employee struggle to offer you relevant products in a big box store? Or, have you ever thought the products being offered to you on a commerce site weren't the most up to date or contextual offerings? These are the types of problems we can address using and exploring linked data. 

    Slide 5: I see a correlation between linked data, business, and the data information knowledge wisdom hierarchy that has been part of information science for a very long time. On the lower tier, we see data and information, two things that we have an abundance of. Where we really want to be is toward the top of the hierarchy. Consumers want to acquire personal knowledge themselves, or be given knowledge by knowledgeable employees in order to make a correct purchasing decision. In order for businesses like Best Buy to remain competitive we need to strive to acquire wisdom from our data in order to drive smarter business decisions. In linked data terms, data is the raw data, information is structuring raw data using vocabularies/ ontologies to provide definition, knowledge is connecting data objects together to form relationships, and wisdom is exploring these connections for insight.  

    Slide 6: I wanted to show you a very basic graph of product relationships using the GoodRelations vocabulary -- one of the strongest and most well defined ontologies for ecommerce. As you can see, products A,B, and C have relationships with each other, connected by the GoodRelations issimilarto method. They also share a relationship with sub product X, which is an add on or accessory. Product C has an additional relationship to sub product X -- with this model we can infer that sub product X may also be an add on to product A and B. Now imagine taking this small, simple model, expanding it using other methods from GoodRelations, and iterating over a product catalog of hundreds of thousands or even millions of products. I see a huge potential for deep product discovery which can benefit both consumers and retailers. 

    Slide 7: After we acquire wisdom from linked data there are many ways to share it with humans and machines. The first thing that comes to my mind is the augmentation of human knowledge with wisdom from Linked Data. Over the past few years customers have identified gaps between human/employee knowledge and knowledge available on the web. I see a potential for bridging this gap by putting human consumable visual data experiences powered by linked data into the hands of every employee, say on a handheld device -- expanding their knowledge of complex products and sharing that knowledge shoulder to shoulder with consumers.

    Enabled with rich insight from Linked data we can also start to build smarter web sites with web services that translate machine readable Linked data into human readable web experiences that are ultra contextual and relevant to the end user.

    Lastly, Best Buy is one of the first retailers to publish structured data to the web for consumption by outside parties, most notably search engines. Recently a group of the major search engines has come together to form a coalition of companies under the schema.org name that support rich data markup like RDFa 1.1 and Microdata in the HTML pages we produce every day, establishing an easy way for everyday HTML authors to contribute to a very rich open web of Linked Data.
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
No Downloads
Views
Total Views
1,581
On Slideshare
0
From Embeds
0
Number of Embeds
2
Actions
Shares
0
Downloads
17
Comments
1
Likes
5
Embeds 0
No embeds

Report content
Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
No notes for slide
  • Thanks for the introduction – I welcome the opportunity to give my perspective on the impact Linked Data can have on retail
  • Products can be complex objects, and the product landscape is vast -- millions if not billions of products available to the average consumer.  Products in the Best Buy catalogue have an average of 31 actionable attributes (attributes that can influence a purchasing decision) ...overall actionable attributes range from a couple,in the case of a simple object like a gift card, to near 100 in the case of a home appliance. While consumers have a literal paradox of choice, a very small percentage of products make up a majority of consumer purchases and company revenue (estimated between 10-15%), leaving 85-90% of product offerings undiscovered or underutilized. This discrepancy is something I’ve called the “product long tail”, where a minority of retail products make up a majority of customer purchases and company revenue.
  • One of the reasons I believe there is such a long tail of under sold products is that we haven't used the set of built in complex relationships that products naturally have with one another to our advantage. Products can be linked together in a number of different ways using linked data to form relationships. These relationships can be explored for the benefit of the consumer and the company. Its a level of product discovery that goes way beyond "people who purchased this item also bought ” scenario that is so common on many retail web sites.  
  • With these complex product offerings and deep product relationships, humans (both consumers and employees) are at a disadvantage. As a consumer have you ever been challenged to find that one product that exactly fits your needs? Or have you experienced an employee struggle to offer you relevant products in a big box store? Or, have you ever thought the products being offered to you on a commerce site weren't the most up to date or contextual offerings? These are the types of problems we can address using and exploring linked data. 
  • I see a correlation between linked data, business, and the data information knowledge wisdom hierarchy that has been part of information science for a very long time. On the lower tier, we see data and information, two things that we have an abundance of. Where we really want to be is toward the top of the hierarchy. Consumers want to acquire personal knowledge themselves, or be given knowledge by knowledgeable employees in order to make a correct purchasing decision. In order for businesses like Best Buy to remain competitive we need to strive to acquire wisdom from our data in order to drive smarter business decisions. In linked data terms, data is the raw data, information is structuring raw data using vocabularies/ ontologies to provide definition, knowledge is connecting data objects together to form relationships, and wisdom is exploring these connections for insight.  
  • I wanted to show you a very basic graph of product relationships using the GoodRelations vocabulary -- one of the strongest and most well defined ontologies for ecommerce. As you can see, products A,B, and C have relationships with each other, connected by the GoodRelationsissimilarto method. They also share a relationship with sub product X, which is an add on or accessory. Product C has an additional relationship to sub product X -- with this model we can infer that sub product X may also be an add on to product A and B. Now imagine taking this small, simple model, expanding it using other methods from GoodRelations, and iterating over a product catalog of hundreds of thousands or even millions of products. I see a huge potential for deep product discovery which can benefit both consumers and retailers. 
  • After we acquire wisdom from linked data there are many ways to share it with humans and machines. The first thing that comes to my mind is the augmentation of human knowledge with wisdom from Linked Data. Over the past few years customers have identified gaps between human/employee knowledge and knowledge available on the web. I see a potential for bridging this gap by putting human consumable visual data experiences powered by linked data into the hands of every employee, say on a handheld device -- expanding their knowledge of complex products and sharing that knowledge shoulder to shoulder with consumers.Enabled with rich insight from Linked data we can also start to build smarter web siteswith web services that translate machine readable Linked data into human readable web experiences that are ultra contextual and relevant to the end user.Lastly, Best Buy is one of the first retailers to publish structured data to the webfor consumption by outside parties, most notably search engines. Recently a group of the major search engines has come together to form a coalitionof companies under the schema.org name that support rich data markup like RDFa 1.1 and Microdata in the HTML pages we produce every day, establishing an easy way for everyday HTML authors to contribute to a very rich open web of Linked Data.
  • Thanks!
  • Transcript

    • 1. Better Retailing Through Linked Data Jay Myers,
    • 2. Products are complex objects Gallon door storage Brand: Samsung Material: Stainless steel Total capacity: 25.8 cu. ft. French doors Sale price: $1,949.99 Regular price: $2,599.99 Height: 68 1/2” Color: BlackModel Number:RF267AERS/XAA Freezer capacity: 8.1 cu. ft. Bottom-loading freezer Depth: 29 1/8” Width: 35 ¾”
    • 3. They also have complex relationships Sub-product Sub-product Sub-product V Sub-product Y U XProduct A Product B Product C Sub-product Sub-product W Z
    • 4. With 523,852 products in my catalog, how many 2, 3, or 4 RELEVANT product combinations can I offer my customer? Where can I find ablack, french door, bottom loading Let’s query the social refrigerator that is web for trendingless than 68” high? product categories, selecting the highest margin products in my catalog to display on the homepage.
    • 5. DIKW, Linked Data, and Business } Companies need to acquire, display wisdom to remain Wisdom competitive Knowledge } Consumer expects personal, employee { knowledge InformationAbundance ofinformation anddata exists Data
    • 6. Sub-product XProduct A gr:isSimilarTo Product B Product C Sub-product X
    • 7. Sharing the Wisdom of Linked DataAugment human Smart web services Publish machine-readableknowledge drive human-readable formats for machines visualizations
    • 8. Thank you!Jay.Myers@bestbuy.co m @jaymyers