People Search

Contents

• Part 1: Intranet people search (case study)

• Part 2: Internet people search

What if Harley had never met Davidson?

VisionVision in the project

What if Bang had never met Olufsen?

Visionen i prosjektet

What can 35000 people do together?

Connected people work smarter

The challengeUndocumented knowledgeUtilizing the weak tiesSharePoint pain points

Challenge: undocumented knowledge

• 80 % of the knowledge is in peoples head *)

*) http://www.cmbg.org/papers/2003Presentations/P04-Capt%20&%20Using%20Val%20Undoc%20Knowl%20-%20Lewis%20Hanes.ppt.

Challenge: undocumented knowledge

• Knowledge workers use as much as 40% of their time to find answers

• They use more than 20% of our time finding answers that others in the company already have

• Less than 20% of available critical knowledge is being used actively

• Employees are more likely to turn to colleagues for information

• Connecting people is the fastest way to speed up knowledge reuse

• Research from Way of Work project

Challenge: utilizing the weak ties

• Collaboration usually done in very small groups– How to make these small groups connect to each other?

– Granovetter’s model of social networks

http://en.wikipedia.org/wiki/Weak_tie

Challenge: SharePoint pain points

Site 1

Site 15

Site 2

Site 1

Site 1000Site 1

Site 345

Site 6

Site 1

Site 1

Site 3

Site 1

Site 4

Site 1

Site 300Site 7

Site nSite 100

Site 46

Blog 3

Blog 1

Site 50

Site 200

Blog 2

Site 1 Site 2

Site 100

Site 100

Site 100

Site 100 Site 100

Site 100Collaboration

Human resources

Recruiting

Challenge: the collaboration cycle

Phase Tools Challenge

Find Search, navigation Content must be searchable/navigable

Choose Navigable search result Good relevance indicators

Communicate IM, email etc Immediate contact

Create relation

Knowledge network Reify relation

Collaborate Wikis, workspaces etc Ease of use, no extra work

Share Blog, CMS, DM etc Ease of use, good metadata

Find Search, navigation Content must be searchable/navigable

The collaboration cycle. Jørgen Dalen and Anne Zanussi

FindabilityFace recognitionThe use of conceptsExploration vs targeted behavior

FindabilityFace recognitionThe use of conceptsExpoloration vs targeted search

Findability: exploration vs targeted behavior

Is a domain expert in

Usability

User experience

UX dept

Tone Terum

Is a

utho

r of

New guidelines for mobile (news story)

Halogen

Is e

mpl

oyed

in

Has

as

sub-

dom

ain

Colleague w

ith

Has as topic

Exploration Targeted behavior: Persons “hubs” in the information space

Is a service delivered by

Usability

User experience

Tone Terum Is o

wne

d by

Halogen

Is e

mpl

oyed

in

has

as s

ub-d

omai

n

Is expert in

has as topic Mobile guidelines (news article)

writ

ten

by

Findability: the use of concepts

Documents

Workspaces

Site x

Persons (My Site)

• Strategy and finance– Business development

– Business strategy

– Business plan

• Market and product– Marketing and branding

– Market strategy

– Market research and analysis

– Product management

• Sales and customer relations– Customer support

– Direct sales

– Etc

Taxonomy (competence areas) Information assetsPortal pages

Portal

Business strategy

Domain experts

Related documents

Workspaces

Ola NormannKari AndersenJørgen Dalen

Lorem ipsum dolor sit amet, odio iusto set diam nonummy carem clarem.

Workspace 1Workspace 2

Document 1Document 2

Document 3Document 4

Topic pagesWorkspace 1Workspace 2

EthicsThe dilemmas of ranking peopleEthicsThe dilemma of ranking people

Solution name

Ethics: three principles for ranking people

• Social distance: higher ranking to people that is similar– Pros: Increased relevance

– Con: The strength of weak ties not necessarily utilized

Ethics: three principles for ranking people

• Social distance: higher ranking to people that is similar– Pros: Increased relevance

– Con: The strength of weak ties not necessarily utilized

• Contribution: higher ranking to people that contribute– Pros: These people tend to be motivated to contribute

– Cons: Might lead to manipulation of ranking

Ethics: three principles for ranking people

• Social distance: higher ranking to people that is similar– Pros: Increased relevance

– Con: The strength of weak ties not necessarily utilized

• Contribution: higher ranking to people that contribute– Pros: These people tend to be motivated to contribute

– Cons: Might lead to manipulation of ranking

• Consumption: higher ranking to people that are often contacted/bookmarked

– Pros: Reflects the actual interest in this person

– Cons: These people might not want to be contacted

Conclusion

Conclusion intranet search

• Vision– Finding the right person can be more important than finding a document

• Interaction– Finding a person is not enough (connecting, relating, collaborating…)

• Information architecture – Use the same taxonomy for both people and content (not two different

systems)

If the world was my intranet…

People search on the internet

Why interest for people search on Internet?

www.halogen.no/kurs-og-artikler/fagartikler/internett-trender-for-2009/

Why interest for people search on Internet?

• The social media war is moving into the search realm

• Search engines must utilize data on social distance to become more effective

• More people than before contribute with content on the web

• Semantic web offer new possibilities

• More organizations see the necessity with a multi-networked approach

Ambiguous:Which Jørgen Dalen?

Outdated

Redunant

Trivial

Sesam personsøk

Disambiguation, but not possible to filter the content on chosen person

Google profile: what is the purpose?

Google profile: what is the purpose?

• Improve data quality?

• Getting monopoly by “owning” the namespace?

• Evolve into a more social search (utilizing social graphs)

• Offer personal SEO?

• Compete with LinkedIn, Facebook and other person profile pages?

• Enrich the use of other Google applications (Google Talk)?

• http://factoryjoe.com/blog/2009/04/15/google-profiles-namespace-lock-in-social-search/

• http://googleblog.blogspot.com/2009/04/search-for-me-on-google.html

Promising ”semantic web-ish” initiatives

• One identity for login on different web services

• Only applicable for login

• Not accessible for other users

• http://openid.net/

Promising ”semantic web-ish” initiatives

• One identity for login on different web services

• Only applicable for login

• Not accessible for other users

• http://openid.net/

• Unique identity for music artists

• Contains other useful metadata as well

• Web page representation of metadata as well

• Accessible for all

• Not human-readable URI

• http://musicbrainz.org/

Promising ”semantic web-ish” initiatives

• One identity for login on different web services

• Only applicable for login

• Not accessible for other users

• http://openid.net/

• Unique identity for music artists

• Contains other useful metadata as well

• Web page representation of metadata as well

• Accessible for all

• Not human-readable URI

• http://musicbrainz.org/

• Restful URI http://google.com/profiles/username/joergen.dalen.norway

• Owned by Google

• Some content avaliable to all on web

Promising ”semantic web-ish” initiatives

• One identity for login on different web services

• Only applicable for login

• Not accessible for other users

• http://openid.net/

• Unique identity for music artists

• Contains other useful metadata as well

• Web page representation of metadata as well

• Accessible for all

• Not human-readable URI

• http://musicbrainz.org/

• Restful URI http://google.com/profiles/username/joergen.dalen.norway

• Owned by Google

• Some content avaliable to all on web

• Very little content publically available

• Available in web interface

• URI reusable, but locked to Facebook (?)

Summary internet people search

• I would like…– Improved people search with semantic web functionality

– Create relationships between content and persons

– Standardized people metadata and profiles

– One public and one private ID for each person

Summary internet people search

• I would like…– Improved people search with semantic web functionality

– Create relationships between content and persons

– Standardized people metadata and profiles

– One public and one private ID for each person

– Less redundancy (duplicated content and services)

Summary internet people search

• I would like…– Improved people search with semantic web functionality

– Create relationships between content and persons

– Standardized people metadata and profiles

– One public and one private ID for each person

– Less redundancy (duplicated content and services)

– Simpler, more transparent privacy settings (not like Facebook)

Summary internet people search

• I would like…– Improved people search with semantic web functionality

– Create relationships between content and persons

– Standardized people metadata and profiles

– One public and one private ID for each person

– Less redundancy (duplicated content and services)

– Simpler, more transparent privacy settings (not like Facebook)

• I wouldn’t like…– Too much personal control on the content on search pages

– Complete removal of trivial content

Recommended reading

Sigvald J. Harryson: Managing know-who based companies A Multinetworked Approach to Knowledge and Innovation Management

Mark Granovetter:The strength of weak tieshttp://www.si.umich.edu/~rfrost/courses/SI110/readings/In_Out_and_Beyond/Granovetter.pdf