SlideShare a Scribd company logo
1 of 4
Download to read offline
Information and information                                                             last time
             systems                                                • we considered some general features
                                                                      of what makes Google successful
             Module 2 –                                             • and you were asked to consider
   successful information systems
                  Ian Ruthven
                office LT12.16
          email: ir@cis.strath.ac.uk

                                 1                                                                   2




       some of your reasons                                                 some of your reasons
• compared to physical shops                                        • good user interface
  – can send products anywhere in world                                – searching, browsing and suggestion facilities
  – site features such as reviews                                      – neutral colours
• compared to other e-commerce sites                                • bad user interface
  – localised shopping (.co.uk, .com, etc)
     • saves costs on shipping, shop in different languages
                                                                       – cluttered, confusing, lots of scrolling
     • but consistent layout across different parts of site and     • reviews on products
       different domains
  – breadth of database and connection to other sellers             • advertising vs. no advertising
     • good: can find things Amazon does not stock directly, used      – can gain revenue (c.f. Amazon mastercard)
       items are cheaper                                               – some groups argued lack of advertising increases trust
     • bad: maybe confusing for people (think buying Amazon, but
       buying from someone else)                                       – confusion over how much advertising Amazon does
  – ease of use (one-click shopping for example)                    • reliability (or at least perceived reliability)
  – security (also ability to track purchases makes site feel       • personalisation
    more secure)
                                                                       – wish lists, recommendations, personalised stores, etc
                                 3                                                                   4




                       Amazon                                                           competitors
                                                                    • www.bookshop.co.uk (Internet
• Amazon’s tasks                                                      Bookshop – pre-Amazon, now WHSmith
  – user task: to allow you to buy things                             owned)
  – developers’ task: to sell you things                            • 3 ways to find
• but lots of sites do this, so why is                              books
  Amazon successful?                                                • search
• start with user tasks                                             • browse
                                                                    • offers

                                 5                                                                   6




                                                                                                                                  1
browse                                            Waterstones (same thing)




                                   7                                                                 8




                          others                                         Amazon (mostly same thing)
• most e-commerce systems offer the same features
• search
   – good for directed finding
   – bad for general finding (what is there, what might I like)
• browse (categories)
   – good for general finding
   – ok for directed finding
      • but can be bad (slow) for directed finding
   – bad for large data sets (too many categories - where is
     something stored, is there something new…)
• special offers
   – small number of specially displayed items
      • good for new items (informs user)
      • good for popular items (saves time)
      • bad for almost everything else

                                   9                                                                10




   several things make Amazon
                                                                                                 also
            successful
• v. large database                                                     • some fun things
   – including out of print and used books
      • not restricted catalogue                                           – like different representations of books
   – not just buying from Amazon but a network of booksellers              – concordance (100 most common words)
      • with reputations like e-bay
      • c.f. Google one answer better than none especially for repeat      – which book?
        business                                                        • again among another away behind ben black boat cam
   – not just books                                                       e captain come cried day dead doctor door down end
                                                                          enough even eyes face far first flint found get go goo
      • one stop shop – web supermarket                                   d got great hand hawkins head heard himself hispanio
• useful features                                                         la house island jim john know last lay left little livese
                                                                          y long look looked man may men might moment mr
   – reviews by readers, readability scores, sample pages                 must nothing now old once own place put right round
   – thinking about what information a consumer might want                  rum saw say sea see seemed seen ship should side s
      • and what might be more difficult about buying things without      ilver sir soon squire still sure take tell thing think ho
        seeing them                                                       ugh thought three till time told took treasure two upo
• nice features                                                           n voice water went word
   – filter searches by price, genre
                                   11                                                               12




                                                                                                                                      2
others                                            more importantly
• statistically improbably phrases                            • also good support
  – north inlet, apple barrel, seafaring man,                   – can return things just because you
    seafaring men                                                 changed your mind
                                                                – wish lists for nice friends to buy you
• capitalized phrases (CAPs):                                     presents
 Ben Gunn, Long John, Captain Smollett, Black Dog, Admiral
 Benbow, John Silver, Jim Hawkins, Skeleton Island, Cap'n     • but more importantly
 Smollett, Billy Bones, Captain Flint, Israel Hands, George     – recommendations
 Merry, Cap'n Silver, Doctor Livesey, Tom Morgan, Cape of
 the Woods, Job Anderson, Tom Redruth, Darby M'Graw,          • uses idea of collaborative filtering
 Davy Jones, Execution Dock, Narrative Continued                – based on similarity of people’s taste
• all can be used as search keys                                   • “If Crawford likes Scooby Doo, I will like
                                                                     Scooby Doo”
                             13                                                           14




        Collaborative filtering
                                                                    Collaborative filtering
             (two levels)
                                                              • collaborative filtering based on user
                                                                profiles tracking
                                                                   • what books/cds/dvds/etc people have bought
                                                                   • what web pages they have visited
                                                                   • what food they buy (supermarkets)
                                                              • aim: to recommend items to people
                                                                who share similar tastes
                                                                   • also called recommender systems



                             15                                                           16




      collaborative filtering                                                     ratings
                                                              • can also include ratings
                                                                – explicit vs implicit
         1      2       3         4    5




          2      12      8        9     7                                    1       2         3    4     5




          1      3       5        4     6                                     2      12        8     9     7

              • has similar tastes to
              • has v. different tastes
                to                                                            1      3         5     4     6
                             17                                                           18




                                                                                                                  3
collaborative filtering                                                          two problems
• suggests things you did not even know                                   • size of data
  you wanted….                                                              – based on most similar person
                                                                               • need large group
   – and therefore you buy more things                                         • need relatively large number of ratings/purchases
     (system goal)                                                        • ‘cold-start’ problem
   – and buy them from Amazon                                               – new items to database
• particularly useful for complex objects                                      • cannot recommend without previous use/purchase
                                                                               • cannot purchase without recommendation
   – how to describe what kind of dvds you like                             – new people
     or what type of music?
                                                                            – usually need some hybrid system
      • easy to recognise harder to search for                                 • querying/browsing for new items, recommendation for
                                                                                 others
                                     19                                                                20




                           Amazon                                                              Amazon
• Amazon, like Google, is trying to capture a task
  market                                                                  • some you did not mention
   – Amazon = buy things                                                    – no physical stock control issues
   – Google = search for things
• tries to make shopping
                                                                            – no costs on physical buildings (in
   – easy                                                                     expensive shopping areas)
      • one-click, delivery to many addresses, gift-wrapping etc
                                                                            – lower costs on staff
   – successful
      • wide database, recommendation                                     • it’s not a real shop!
   – happen more often
      • recommendation, wish-lists                                          – competition is not just online
   – lightweight
      • fun things to keep you interested and interacting with the site
            – so therefore you buy more things!!!!


                                     21                                                                22




                          next time                                                          next time
• you have a choice between two new-                                      • you have a choice between two new-
  ish but popular sites….                                                   ish but popular sites….




                                     23                                                                24




                                                                                                                                       4

More Related Content

Similar to How Amazon uses collaborative filtering to provide personalized recommendations

Mechanical Librarian
Mechanical LibrarianMechanical Librarian
Mechanical LibrarianAndre Vellino
 
Yvonne Doll, Designing Content for Usability
Yvonne Doll, Designing Content for UsabilityYvonne Doll, Designing Content for Usability
Yvonne Doll, Designing Content for Usabilitywebcontent2007
 
What is the Semantic Web
What is the Semantic WebWhat is the Semantic Web
What is the Semantic WebIvan Herman
 
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09BookNet Canada
 
Digital preservation and institutional repositories
Digital preservation and institutional repositoriesDigital preservation and institutional repositories
Digital preservation and institutional repositoriesDorothea Salo
 
NAP review
NAP reviewNAP review
NAP reviewpducy
 
PS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommercePS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommerceIan Jindal
 
PS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommercePS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommerceguest904b2b
 
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade Insights
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade InsightsAIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade Insights
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade InsightsAIPMM Administration
 
Bibliographic Database Integrity
Bibliographic Database IntegrityBibliographic Database Integrity
Bibliographic Database IntegrityEvergreen ILS
 
Extreme Makeover: Web Site Edition (OPLIN)
Extreme Makeover: Web Site Edition (OPLIN)Extreme Makeover: Web Site Edition (OPLIN)
Extreme Makeover: Web Site Edition (OPLIN)Laura Solomon
 
The Art of The Start
The Art of The StartThe Art of The Start
The Art of The Startelliekem
 
Wading Through the Web
Wading Through the WebWading Through the Web
Wading Through the Webkjhatzi
 
Educational Uses of Web 2.0 Based Applications with Notes
Educational Uses of Web 2.0 Based Applications with NotesEducational Uses of Web 2.0 Based Applications with Notes
Educational Uses of Web 2.0 Based Applications with NotesMike Qaissaunee
 
Web 2.0 Managerial Economics
Web 2.0 Managerial EconomicsWeb 2.0 Managerial Economics
Web 2.0 Managerial EconomicsAvinash Singh
 

Similar to How Amazon uses collaborative filtering to provide personalized recommendations (19)

Mechanical Librarian
Mechanical LibrarianMechanical Librarian
Mechanical Librarian
 
Yvonne Doll, Designing Content for Usability
Yvonne Doll, Designing Content for UsabilityYvonne Doll, Designing Content for Usability
Yvonne Doll, Designing Content for Usability
 
Irl Web Strategy
Irl Web StrategyIrl Web Strategy
Irl Web Strategy
 
eBags May 2008
eBags May 2008eBags May 2008
eBags May 2008
 
What is the Semantic Web
What is the Semantic WebWhat is the Semantic Web
What is the Semantic Web
 
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09
Shortcovers - Michael Serbinis - BookNet Canada Tech Forum 09
 
Digital preservation and institutional repositories
Digital preservation and institutional repositoriesDigital preservation and institutional repositories
Digital preservation and institutional repositories
 
NAP review
NAP reviewNAP review
NAP review
 
PS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommercePS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommerce
 
PS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommercePS067-bazaarvoice-socialcommerce
PS067-bazaarvoice-socialcommerce
 
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade Insights
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade InsightsAIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade Insights
AIPMM Webcast: Framework Fight Club with Sean Campbell of Cascade Insights
 
Bibliographic Database Integrity
Bibliographic Database IntegrityBibliographic Database Integrity
Bibliographic Database Integrity
 
Extreme Makeover: Web Site Edition
Extreme Makeover: Web Site EditionExtreme Makeover: Web Site Edition
Extreme Makeover: Web Site Edition
 
Extreme Makeover: Web Site Edition (OPLIN)
Extreme Makeover: Web Site Edition (OPLIN)Extreme Makeover: Web Site Edition (OPLIN)
Extreme Makeover: Web Site Edition (OPLIN)
 
The Art of The Start
The Art of The StartThe Art of The Start
The Art of The Start
 
Wading Through the Web
Wading Through the WebWading Through the Web
Wading Through the Web
 
Educational Uses of Web 2.0 Based Applications with Notes
Educational Uses of Web 2.0 Based Applications with NotesEducational Uses of Web 2.0 Based Applications with Notes
Educational Uses of Web 2.0 Based Applications with Notes
 
Search Engine Google
Search Engine GoogleSearch Engine Google
Search Engine Google
 
Web 2.0 Managerial Economics
Web 2.0 Managerial EconomicsWeb 2.0 Managerial Economics
Web 2.0 Managerial Economics
 

Recently uploaded

Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdf
Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdfCatalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdf
Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdfOrient Homes
 
0183760ssssssssssssssssssssssssssss00101011 (27).pdf
0183760ssssssssssssssssssssssssssss00101011 (27).pdf0183760ssssssssssssssssssssssssssss00101011 (27).pdf
0183760ssssssssssssssssssssssssssss00101011 (27).pdfRenandantas16
 
BEST ✨ Call Girls In Indirapuram Ghaziabad ✔️ 9871031762 ✔️ Escorts Service...
BEST ✨ Call Girls In  Indirapuram Ghaziabad  ✔️ 9871031762 ✔️ Escorts Service...BEST ✨ Call Girls In  Indirapuram Ghaziabad  ✔️ 9871031762 ✔️ Escorts Service...
BEST ✨ Call Girls In Indirapuram Ghaziabad ✔️ 9871031762 ✔️ Escorts Service...noida100girls
 
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...Dave Litwiller
 
GD Birla and his contribution in management
GD Birla and his contribution in managementGD Birla and his contribution in management
GD Birla and his contribution in managementchhavia330
 
Progress Report - Oracle Database Analyst Summit
Progress  Report - Oracle Database Analyst SummitProgress  Report - Oracle Database Analyst Summit
Progress Report - Oracle Database Analyst SummitHolger Mueller
 
Regression analysis: Simple Linear Regression Multiple Linear Regression
Regression analysis:  Simple Linear Regression Multiple Linear RegressionRegression analysis:  Simple Linear Regression Multiple Linear Regression
Regression analysis: Simple Linear Regression Multiple Linear RegressionRavindra Nath Shukla
 
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRL
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRLMONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRL
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRLSeo
 
Insurers' journeys to build a mastery in the IoT usage
Insurers' journeys to build a mastery in the IoT usageInsurers' journeys to build a mastery in the IoT usage
Insurers' journeys to build a mastery in the IoT usageMatteo Carbone
 
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999Tina Ji
 
The CMO Survey - Highlights and Insights Report - Spring 2024
The CMO Survey - Highlights and Insights Report - Spring 2024The CMO Survey - Highlights and Insights Report - Spring 2024
The CMO Survey - Highlights and Insights Report - Spring 2024christinemoorman
 
Ensure the security of your HCL environment by applying the Zero Trust princi...
Ensure the security of your HCL environment by applying the Zero Trust princi...Ensure the security of your HCL environment by applying the Zero Trust princi...
Ensure the security of your HCL environment by applying the Zero Trust princi...Roland Driesen
 
DEPED Work From Home WORKWEEK-PLAN.docx
DEPED Work From Home  WORKWEEK-PLAN.docxDEPED Work From Home  WORKWEEK-PLAN.docx
DEPED Work From Home WORKWEEK-PLAN.docxRodelinaLaud
 
Tech Startup Growth Hacking 101 - Basics on Growth Marketing
Tech Startup Growth Hacking 101  - Basics on Growth MarketingTech Startup Growth Hacking 101  - Basics on Growth Marketing
Tech Startup Growth Hacking 101 - Basics on Growth MarketingShawn Pang
 
Eni 2024 1Q Results - 24.04.24 business.
Eni 2024 1Q Results - 24.04.24 business.Eni 2024 1Q Results - 24.04.24 business.
Eni 2024 1Q Results - 24.04.24 business.Eni
 
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...Dipal Arora
 
It will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 MayIt will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 MayNZSG
 
Monthly Social Media Update April 2024 pptx.pptx
Monthly Social Media Update April 2024 pptx.pptxMonthly Social Media Update April 2024 pptx.pptx
Monthly Social Media Update April 2024 pptx.pptxAndy Lambert
 

Recently uploaded (20)

Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdf
Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdfCatalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdf
Catalogue ONG NƯỚC uPVC - HDPE DE NHAT.pdf
 
0183760ssssssssssssssssssssssssssss00101011 (27).pdf
0183760ssssssssssssssssssssssssssss00101011 (27).pdf0183760ssssssssssssssssssssssssssss00101011 (27).pdf
0183760ssssssssssssssssssssssssssss00101011 (27).pdf
 
Nepali Escort Girl Kakori \ 9548273370 Indian Call Girls Service Lucknow ₹,9517
Nepali Escort Girl Kakori \ 9548273370 Indian Call Girls Service Lucknow ₹,9517Nepali Escort Girl Kakori \ 9548273370 Indian Call Girls Service Lucknow ₹,9517
Nepali Escort Girl Kakori \ 9548273370 Indian Call Girls Service Lucknow ₹,9517
 
BEST ✨ Call Girls In Indirapuram Ghaziabad ✔️ 9871031762 ✔️ Escorts Service...
BEST ✨ Call Girls In  Indirapuram Ghaziabad  ✔️ 9871031762 ✔️ Escorts Service...BEST ✨ Call Girls In  Indirapuram Ghaziabad  ✔️ 9871031762 ✔️ Escorts Service...
BEST ✨ Call Girls In Indirapuram Ghaziabad ✔️ 9871031762 ✔️ Escorts Service...
 
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
Enhancing and Restoring Safety & Quality Cultures - Dave Litwiller - May 2024...
 
GD Birla and his contribution in management
GD Birla and his contribution in managementGD Birla and his contribution in management
GD Birla and his contribution in management
 
Progress Report - Oracle Database Analyst Summit
Progress  Report - Oracle Database Analyst SummitProgress  Report - Oracle Database Analyst Summit
Progress Report - Oracle Database Analyst Summit
 
Regression analysis: Simple Linear Regression Multiple Linear Regression
Regression analysis:  Simple Linear Regression Multiple Linear RegressionRegression analysis:  Simple Linear Regression Multiple Linear Regression
Regression analysis: Simple Linear Regression Multiple Linear Regression
 
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRL
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRLMONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRL
MONA 98765-12871 CALL GIRLS IN LUDHIANA LUDHIANA CALL GIRL
 
Insurers' journeys to build a mastery in the IoT usage
Insurers' journeys to build a mastery in the IoT usageInsurers' journeys to build a mastery in the IoT usage
Insurers' journeys to build a mastery in the IoT usage
 
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999
Russian Faridabad Call Girls(Badarpur) : ☎ 8168257667, @4999
 
The CMO Survey - Highlights and Insights Report - Spring 2024
The CMO Survey - Highlights and Insights Report - Spring 2024The CMO Survey - Highlights and Insights Report - Spring 2024
The CMO Survey - Highlights and Insights Report - Spring 2024
 
Ensure the security of your HCL environment by applying the Zero Trust princi...
Ensure the security of your HCL environment by applying the Zero Trust princi...Ensure the security of your HCL environment by applying the Zero Trust princi...
Ensure the security of your HCL environment by applying the Zero Trust princi...
 
Best Practices for Implementing an External Recruiting Partnership
Best Practices for Implementing an External Recruiting PartnershipBest Practices for Implementing an External Recruiting Partnership
Best Practices for Implementing an External Recruiting Partnership
 
DEPED Work From Home WORKWEEK-PLAN.docx
DEPED Work From Home  WORKWEEK-PLAN.docxDEPED Work From Home  WORKWEEK-PLAN.docx
DEPED Work From Home WORKWEEK-PLAN.docx
 
Tech Startup Growth Hacking 101 - Basics on Growth Marketing
Tech Startup Growth Hacking 101  - Basics on Growth MarketingTech Startup Growth Hacking 101  - Basics on Growth Marketing
Tech Startup Growth Hacking 101 - Basics on Growth Marketing
 
Eni 2024 1Q Results - 24.04.24 business.
Eni 2024 1Q Results - 24.04.24 business.Eni 2024 1Q Results - 24.04.24 business.
Eni 2024 1Q Results - 24.04.24 business.
 
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...
Call Girls Navi Mumbai Just Call 9907093804 Top Class Call Girl Service Avail...
 
It will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 MayIt will be International Nurses' Day on 12 May
It will be International Nurses' Day on 12 May
 
Monthly Social Media Update April 2024 pptx.pptx
Monthly Social Media Update April 2024 pptx.pptxMonthly Social Media Update April 2024 pptx.pptx
Monthly Social Media Update April 2024 pptx.pptx
 

How Amazon uses collaborative filtering to provide personalized recommendations

  • 1. Information and information last time systems • we considered some general features of what makes Google successful Module 2 – • and you were asked to consider successful information systems Ian Ruthven office LT12.16 email: ir@cis.strath.ac.uk 1 2 some of your reasons some of your reasons • compared to physical shops • good user interface – can send products anywhere in world – searching, browsing and suggestion facilities – site features such as reviews – neutral colours • compared to other e-commerce sites • bad user interface – localised shopping (.co.uk, .com, etc) • saves costs on shipping, shop in different languages – cluttered, confusing, lots of scrolling • but consistent layout across different parts of site and • reviews on products different domains – breadth of database and connection to other sellers • advertising vs. no advertising • good: can find things Amazon does not stock directly, used – can gain revenue (c.f. Amazon mastercard) items are cheaper – some groups argued lack of advertising increases trust • bad: maybe confusing for people (think buying Amazon, but buying from someone else) – confusion over how much advertising Amazon does – ease of use (one-click shopping for example) • reliability (or at least perceived reliability) – security (also ability to track purchases makes site feel • personalisation more secure) – wish lists, recommendations, personalised stores, etc 3 4 Amazon competitors • www.bookshop.co.uk (Internet • Amazon’s tasks Bookshop – pre-Amazon, now WHSmith – user task: to allow you to buy things owned) – developers’ task: to sell you things • 3 ways to find • but lots of sites do this, so why is books Amazon successful? • search • start with user tasks • browse • offers 5 6 1
  • 2. browse Waterstones (same thing) 7 8 others Amazon (mostly same thing) • most e-commerce systems offer the same features • search – good for directed finding – bad for general finding (what is there, what might I like) • browse (categories) – good for general finding – ok for directed finding • but can be bad (slow) for directed finding – bad for large data sets (too many categories - where is something stored, is there something new…) • special offers – small number of specially displayed items • good for new items (informs user) • good for popular items (saves time) • bad for almost everything else 9 10 several things make Amazon also successful • v. large database • some fun things – including out of print and used books • not restricted catalogue – like different representations of books – not just buying from Amazon but a network of booksellers – concordance (100 most common words) • with reputations like e-bay • c.f. Google one answer better than none especially for repeat – which book? business • again among another away behind ben black boat cam – not just books e captain come cried day dead doctor door down end enough even eyes face far first flint found get go goo • one stop shop – web supermarket d got great hand hawkins head heard himself hispanio • useful features la house island jim john know last lay left little livese y long look looked man may men might moment mr – reviews by readers, readability scores, sample pages must nothing now old once own place put right round – thinking about what information a consumer might want rum saw say sea see seemed seen ship should side s • and what might be more difficult about buying things without ilver sir soon squire still sure take tell thing think ho seeing them ugh thought three till time told took treasure two upo • nice features n voice water went word – filter searches by price, genre 11 12 2
  • 3. others more importantly • statistically improbably phrases • also good support – north inlet, apple barrel, seafaring man, – can return things just because you seafaring men changed your mind – wish lists for nice friends to buy you • capitalized phrases (CAPs): presents Ben Gunn, Long John, Captain Smollett, Black Dog, Admiral Benbow, John Silver, Jim Hawkins, Skeleton Island, Cap'n • but more importantly Smollett, Billy Bones, Captain Flint, Israel Hands, George – recommendations Merry, Cap'n Silver, Doctor Livesey, Tom Morgan, Cape of the Woods, Job Anderson, Tom Redruth, Darby M'Graw, • uses idea of collaborative filtering Davy Jones, Execution Dock, Narrative Continued – based on similarity of people’s taste • all can be used as search keys • “If Crawford likes Scooby Doo, I will like Scooby Doo” 13 14 Collaborative filtering Collaborative filtering (two levels) • collaborative filtering based on user profiles tracking • what books/cds/dvds/etc people have bought • what web pages they have visited • what food they buy (supermarkets) • aim: to recommend items to people who share similar tastes • also called recommender systems 15 16 collaborative filtering ratings • can also include ratings – explicit vs implicit 1 2 3 4 5 2 12 8 9 7 1 2 3 4 5 1 3 5 4 6 2 12 8 9 7 • has similar tastes to • has v. different tastes to 1 3 5 4 6 17 18 3
  • 4. collaborative filtering two problems • suggests things you did not even know • size of data you wanted…. – based on most similar person • need large group – and therefore you buy more things • need relatively large number of ratings/purchases (system goal) • ‘cold-start’ problem – and buy them from Amazon – new items to database • particularly useful for complex objects • cannot recommend without previous use/purchase • cannot purchase without recommendation – how to describe what kind of dvds you like – new people or what type of music? – usually need some hybrid system • easy to recognise harder to search for • querying/browsing for new items, recommendation for others 19 20 Amazon Amazon • Amazon, like Google, is trying to capture a task market • some you did not mention – Amazon = buy things – no physical stock control issues – Google = search for things • tries to make shopping – no costs on physical buildings (in – easy expensive shopping areas) • one-click, delivery to many addresses, gift-wrapping etc – lower costs on staff – successful • wide database, recommendation • it’s not a real shop! – happen more often • recommendation, wish-lists – competition is not just online – lightweight • fun things to keep you interested and interacting with the site – so therefore you buy more things!!!! 21 22 next time next time • you have a choice between two new- • you have a choice between two new- ish but popular sites…. ish but popular sites…. 23 24 4