Putting the Person into Personalization by Elizabeth Churchill
1 like703 views
This document discusses personalization and how it can be improved by putting people at the center. It notes that personalization is approached differently in marketing, engineering, and design. It reviews relevant areas like human-computer interaction, algorithmic living, and various disciplines that can inform personalization like social sciences, machine learning, and data sciences. It advocates designing systems with an awareness of data and collecting data with an awareness of its design rationale and purpose. It also discusses personalizing search outcomes, recommendations, and consumption experiences while considering both process and outcomes as well as moving beyond a generic view of consumers.
Putting the Person into Personalization by Elizabeth Churchill
- 1. Pu#ng the Person into Personaliza1on Elizabeth F. Churchill eBay Research Labs June 2014
- 2. U=lity noun: u=lity -‐ the state of being useful, profitable, or beneficial -‐ a public u=lity adjec=ve: u=lity -‐ useful, able to perform several func=ons -‐ func=onal rather than aJrac=ve Infrastructural
- 3. Changing models of commerce • Consump=on • Distribu=on • Produc=on • “Personaliza=on” ….. different perspec=ves – Marke=ng, content & product targe=ng – Engineering as customisa=on – Design as co-‐produc=on
- 4. Some context
- 5. ψ
- 6. Human Computer Interac=on a discipline concerned with the design, implementa3on and evalua3on of interac3ve compu3ng systems for human use and with the study of major phenomena surrounding them
- 7. Algorithmic Living*(s) Behaviour system – rela=ons of the data Source System Data system – interfaces are conduit from source to data “Algorithmic Living” is a phrase borrowed from Paul Dourish (UCI) and colleagues
- 8. Social Sciences Design Sciences Computa=onal Sciences Design paEerns Informa3on (re)representa3on Informa3on architecture, interac3on design, graphic design Cross device interac3ons Field research methods Prototyping Human Factors Experience design Brain & Body Sciences Percep3on, emo3on Cogni3ve Psychology Decision making (interrup3on and), memory, language, emo3on, individual differences Social psychology Social iden3ty, trust, communica3on Anthropology Shopping & the circula3on of goods Machine Learning Personaliza3on/Recommenda3on Informa3on architecture(s) Catalogues, user representations Search Sciences Data Sciences, Social Network Analysis & Computa3onal sociology Sociology Demographics, SNA Web services Graph theory Collabora3ve Filtering Collabora3ve Filtering Data mining Design recommenda3ons & requirements Data Data representa3on Adap3ve interfaces & interac3ions System requirements HCI & Product/ Service Design Computer Graphics Opera3ng Systems Linguis3cs
- 10. Design aware data, data aware design • The interface brokers a conversa=on between a user and a service, it invites (“affords”) ac=ons • Proac=ve data collec=on versus reac=ve data analy=cs – Too much focus on the numbers and not on the insights – Ar=factual collec=ons versus inten=onal programma=c collec=on – Assume • Tangible and intangible signals • Weak signals • Par=al footprints, composite persons (shared devices/accounts), bots • Misleading informa=on, missing pieces, incomplete stories • Ques=ons to ask yourself – What are you asking to know about a person? Why? • What is the design ra=onale for your instrumenta=on? – Are the data you are gathering fit for your purpose, valid, reliable? • Are your conclusions jus=fied?
- 11. Data shaping, data design People - ”users” Teams Organisational Societal/Regulatory Interfaces Micro Macro & Meta Business Meso
- 12. People - ”users” Teams Organisational Societal/Regulatory Interfaces Micro Macro & Meta Business Meso Experience mining For “deep understanding” rather than “deep learning”, we need to triangulate methods: experiments, ethnography, interviews
- 13. Triangulate data Triangulate perspec3ves Explore different units of analysis Eyetracking Lab experiments Interviews Prototypes (lo and hi) Focus groups Surveys Ethnographic field studies Experimental field studies Ac=vity log (”trace”) analysis Data mining Data visualiza=on Computer scien=sts Anthropologists Sociologists Vision Scien=sts Designers: product, interac=on, graphic
- 14. Personaliza1on Increasing Relevance of Content and Presenta=on Modality/Style
- 15. Recommender Systems – the abstracted & “generic” consumer
- 16. Concierge personaliza=on: “the ‘n’ of 1”
- 18. Outcome Personaliza=on: Search Non-‐personalized Personalized F(f)ossil collector E-‐commerce Product Search: Personaliza3on, Diversifica3on, and Beyond, A=sh Das Sarma, Nish Parikh, Neel Sundaresan, Tutorial at WWW-‐2014
- 20. outcome what (e.g., content match/relevance) process how, where, when
- 21. Kinect Myo Google glass Emo=v, brain & eye control Process
- 22. Consump1on
- 23. Auc=onweb …the history of eBay hJp://mashable.com/2010/08/07/ebay-‐facts/
- 27. Value is beyond the ar=fact The broken laser pointer…the history of eBay hJp://mashable.com/2010/08/07/ebay-‐facts/
- 32. Literacy, reputa=on & trust
- 37. Commerce 3.0 for Development: The promise of the Global Empowerment Network
- 38. Distribu1on
- 39. Commerce 3.0 for Development: The promise of the Global Empowerment Network
- 40. Commerce 3.0 for Development: The promise of the Global Empowerment Network
- 41. Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping
- 42. Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping
- 43. Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping
- 44. Commerce 3.0 for Development: The promise of the Global Empowerment Network
- 45. Digital consump1on is s1ll about material produc1on & distribu1on, global s1ll means local delivery or pick-‐up
- 46. Digital consump1on is s1ll about material produc1on & distribu1on, global s1ll means local delivery or pick-‐up
- 47. Digital consump1on is s1ll about material produc1on & distribu1on, global s1ll means local delivery or pick-‐up
- 48. Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping Under the percentages are people and their prac1ces….
- 49. Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping Under the percentages are people and their purchases….
- 50. Produc1on Changing modes of
- 56. Crowdfunding and “par1cipatory” produc1on
- 57. Made to order – “possibility” shop fronts
- 58. Printed to order – “possibility” shop fronts
- 59. Printed to order – “possibility” shop fronts
- 62. Algorithmic Living(s) Behaviour system – rela=ons of the data Source System Data system – interfaces are conduit from source to data
- 63. Changing models of commerce • Consump=on • Distribu=on • Produc=on • “Personaliza=on” ….. different perspec=ves – Marke=ng, content & product targe=ng – Engineering as customisa=on – Design as co-‐produc=on
- 64. Commerce 3.0 for Development: The promise of the Global Empowerment Network
- 65. (David) Ayman Shamma M. Cameron Jones Elizabeth F. Churchill Ques=ons &/or Comments? churchill@acm.org
- 66. References • Pung the person back into personaliza3on. interac=ons 20.5 (2013): 12-‐15, Elizabeth Churchill • From data divina3on to data-‐aware design. interac=ons 19.5 (2012): 10-‐13, Elizabeth Churchill • E-‐commerce Personaliza3on at Scale 23rd Interna=onal Conference on Informa=on and Knowledge Management (CIKM), Elizabeth Churchill, A=sh Das Sarma, Ranjan Sinha • Data Design for Personaliza3on: Current Challenges and Emerging Opportuni3es, Elizabeth Churchill, A=sh Das Sarma, Workshop at WSDM-‐2014 • E-‐commerce Product Search: Personaliza3on, Diversifica3on, and Beyond, A=sh Das Sarma, Nish Parikh, Neel Sundaresan, Tutorial at WWW-‐2014 • Commerce 3.0 for Development: The promise of the Global Empowerment Network hJp://www.ebaymainstreet.com/news-‐events/commerce-‐30-‐ development-‐promise-‐global-‐empowerment-‐network • Modern Spice Routes: The Cultural Impact of Cross-‐Border Shopping hJps:// www.paypal-‐media.com/assets/pdf/fact_sheet/ PayPal_ModernSpiceRoutes_Report_Final.pdf