Unstructured Content Management

Taxonomic Publishing Models

Tom Reamy Chief Knowledge Architect KAPS Group Knowledge Architecture Professional Services http://www.kapsgroup.com [email protected]

Agenda
 KAPS Group and Knowledge Architecture  Current State of Content Management  Taxonomies and Content Management

What, Why, and How

 Taxonomic Content Management  Infrastructure Content Management

Technology, Teams, Taxonomies Beyond Taxonomies

Knowledge Objects, Semantic Web, Personas, etc.

KAPS Group
 KAPS Background

Knowledge Architecture Consultants Intellectual Infrastructure: Content, Tools, People

E-Learning and Information Architecture Knowledge Architecture Audit Social Network Analysis and Business Process

Professional Services partner to Search, Content Management & Categorization Companies

Current State of Content Management

 Forrester Research:

Current content management systems are immature. High Cost, Proprietary technology Poor implementation, Difficult to maintain & customize Software, system integration Version control, work flow Decoupling content and presentation Delphi Survey
68% finding information is difficult 50% spend more than 2 hours a day looking

 CM is good at:

 Part of a Broader problem

Current State of Content Management

 Content Management

Strong on management, weak on content Content is a black box simply moved around In-depth and articulated understanding of content 90% plan on taxonomy strategy in 24 months 76% taxonomy is important

 What is missing is the meaning dimension

 Perceived Solution Delphi Survey Taxonomy

 Taxonomy: necessary but not sufficient

Content Management and Taxonomy: What?

 Formal Taxonomies

Linnaeus Taxonomy of life Only relationship is Is Kind Of Yahoo Hierarchical Variety of relationships

 Browse Taxonomies (Informal)

 Classifications and Categorization  Metaphorical Taxonomies

Thesaurus, catalog, index, site map

Content Management and Taxonomy: What?

 What makes a good taxonomy?

Formal: Quality Metrics

Corpus, Coverage, Nomenclature, terminology, dependency Mixed classes, verbal forms, bad speciation, etc. Bell Curve, balance of depth and width

Informal: An understandable organization of content that enables people to find information and which supports knowledge discovery.
Creates a context within which facts are related Find, Identify, Describe information, relations, context

Content Management and Taxonomy: What?

 Taxonomy as part of knowledge organization  Metadata: Dublin Core+

CM functions: Language, Identifier, Rights Combination functions: Publisher, Author Subject matter functions: keywords, descriptions
Minimum need controlled vocabularies

 Contextual

DocumentObjectType, AudienceType People, Companies, Compounds, Geography Multiple views into content Dynamic Mapping of facets

 Facets and entities

Content Management and Taxonomy: Why?

 Search Stinks

Integrated Browse and Search works better than search

Ecommerce 56% of all searches fail = lost income Intranet = lost time, lost business, lost ideas Taxonomic CM - Rich semantic web of concepts, not a unstructured collection of documents

 Cost of poor Search and Content Management

If its not organized,you cant find it. If you cant find it, you cant use it. If you cant find it, you waste a lot of time. If you cant find it, you could lose an account. If you cant find it, you could look stupid. If you cant find it, it doesnt exist.

Content Management and Taxonomy: Why?

 In 2 years, categorization will replace search  Categorization will be a component/foundation for:

Search, content management, portals, CRM, collaboration, etc.

 Beyond browse
Agent profiles just in time news Intelligent agents semantic web Contextualized search results Personalization within communities

10

Content Management and Taxonomy: How?

 Old Answer: Manual
hire a bunch of librarians and IAs Costly, difficult to maintain Use SMEs Costly, difficult to maintain, bad track record

 New Answer:
Integrate Manual and new software Integrate Content Management and Taxonomy Integrate central team and local authors

11

Content Management and Taxonomy: How?

 New Technologies

Unstructured Data Management Taxonomy Management Smart Categorization, summarization Entity Extraction and metadata generation Visualization of taxonomic relationships Linguistic analysis, not just bag of words

12

Machine-Categorization: Methods
 Semi-Automatic: Rules, If-Then
Maximum precision & flexibility

 Catalog by Example: Bayesian, SVM, Neural

Training Sets (5-500) Speed, Learning

 Statistical Clustering

Set of Documents & Taxonomy Level

 Semantic Analysis & World Knowledge

13

Machine Categorization: The Human Element

 Automatic Categorization is Not  Humans are better, but not as consistent

Bring outside contexts to the document

Purpose, similar documents, common sense

Understandable mistakes

 Computers are faster and cheaper  Categorization is part of knowledge organization

Meta data, communities, taxonomies, etc.

 The Best Answer is Hybrid or Cyborg Categorization

Taxonomic Content Management: Standard View

Site Owner Pulls Files from Staging Server Makes Edits and Changes Send files to staging server Central Team

Test Files (QA)

Move Files to PreProd Move to Production

15

Send email request to CT to review

Taxonomic Content Management: Taxonomic View

Authors Pulls Files from Staging Server Makes Edits and Changes Send files to staging server Categorization MetaData Taxonomies: Content, Communities, Tasks Central Team

Test Files (QA) Categorization MetaData Move Files to PreProd Move to Production
16

Send email request to CT to review

Taxonomic Content Management: Work Flow with Meaning  Preliminary Foundation Work

Design the ontology Develop taxonomies Design metadata standards Collaborative development of controlled vocabularies Have a summary either written by human or software List of metadata suggestions, entities people, places, etc. Provisional categorization Decision: publish or submit for review, central team or community of experts. Request for additional keywords or categorization issues

 Authors, SMEs check document in:

17

Taxonomic Content Management: Work Flow with Meaning

 Central Team

Review documents easier, faster Use summaries, metadata, entities to provide context Review infrastructure requests new keywords, categories Strengths of local and central Variety of roles, flexible (few dedicated roles needed). Collaborative categorization and keywords by SME, software, and central team
SMEs can function as central team

 Integrated Work Flow

18

Taxonomic Content Management: Work Flow with Meaning  Publish by Category, not web site

Web site is a terrible unit of organization of content 10 to 10,000 documents Who published is only one dimension Collaboration supported across organization Dynamically generate views, facets, web sites Supports intelligent personalization
Requires metadata to go beyond idiosyncratic views of content

 Flexible & Intelligent Publishing

Prompt on unusual connections

Pre-existing, categories Regulatory or legal issue

19

Taxonomic Content Management: Work Flow with Meaning

 Content Reorganization

Category + Publisher = related document sets Rich web of related content

Content + background contexts Legal/Policy contexts Technical contexts Customer / Task contexts

Support browse by topic, type, task, entity, facets

20

Taxonomic Content Management: Work Flow with Meaning  Design even more important

Taxonomic effort Balance of pre-defined and dynamic Broader context of content, communities, processes Metadata, categorization, summarization, etc. Article EContent October KM and E-Learning CM, LCMS, LMS, KM platform Collaboration: E-Room, Intraspect Search and Portals: Epicentric

 CM companies are developing or buying taxonomic capabilities  CM as a platform technology  CM: Beyond Categorization  CM as part of Intellectual Infrastructure
21

Infrastructure Content Management Technology, Teams, Taxonomies

 Technology

CM -- Least important unless you get it wrong Taxonomic Software

Support articulation of intellectual infrastructure Integrated with CM supports maintenance

KM Platform CM in Context
Search, Portals, Collaboration Supports application of the intellectual infrastructure

22

Infrastructure Content Management Technology, Teams, Taxonomies

 Teams Where?  Best: Central, Dedicated Department

Cross Organizational, Multidisciplinary Practical, real world input

 Part Time, Distributed SMEs, Business owners

 Partners: IT, HR, Corporate Communication, Library, Training  Worst: IT Project Manger, Intranet programming team

23

Infrastructure Content Management Technology, Teams, Taxonomies

       

Teams Who? Knowledge Architect and Learning Object Designers Knowledge Engineers and curriculum developers Knowledge Facilitators and Trainers IT, Web developers, application programmers Librarians and information architects Business analysts and project managers Corporate Communication writers and editors

24

Infrastructure Content Management Technology, Teams, Taxonomies  Teams What?  Infrastructure Activities

Integrate taxonomy across the company

Content, communities, activities

Grow and Develop taxonomies

Taxonomy metrics require skill to fix problems

Design content repositories, update and adapt categorization Package knowledge into K objects, combine with stories, learning histories Metrics and Measurement analyze and enhance Knowledge Architecture Audit Cognitive Difference Geography of Thought
Panda, monkey, banana

25

Infrastructure Content Management Technology, Teams, Taxonomies  Taxonomies and Beyond:  Intellectual infrastructure Context for CM  Taxonomy of Communities

Map of formal and informal communities Social Network Analysis, Personas Community specific vocabularies Integrate with knowledge objects, metadata

 Expertise Location, mentoring, story telling  Communities of Practice  Training

Embedded Learning - Just-In-Time, Performance Support

26

Infrastructure Content Management Technology, Teams, Taxonomies

 Taxonomies and Beyond:  Document not the best unit of organization in all situations  Learning/Knowledge Objects

Chunks of content and XML metadata Reusable, flexible, answer machines Important of context rules for relating objects SCORM+
Semantic Density, typical learning time

 Advanced MetaData

RDF and Semantic Web

subClassOf, seeAlso, isRelatedTo
27

ICM and Applications: Contextualizing Content

 Knowledge Creating

Innovation, E-learning, LMS Collaboration

Distributed Categorization, Community Vocabularies

 Knowledge Sharing / Transmission

Collaboration, Retrieval content and experts Smart Applications, Portals CRM, Data warehouse, text mining, business intelligence

 Knowledge Using

28

ICM and Applications: Contextualizing Content

 Knowledge = information + contexts  Contexts are what gives depth and meaning to information

Let me tell you a story

 Contextualizing Content
Related topics, contexts, content types Rules for relating, integrating contexts

29

Knowledge Retrieval: Contexts

 Search for product name
List of documents that are explicitly about the product Category Views

Features of the product Product comparisons Legal or policy documents

Background Resources
List of Experts, communities Glossaries, internal libraries

30

Knowledge Retrieval: Contexts

 Search for product name
Search and Browse options Text or visual options Offers a variety of contexts:

Related content, best bests (community based and input form central team)

Learns from my behavior and community behavior Usage Analytics based on meaning, not counting clicks

31

Knowledge Retrieval: Contexts

 Search for product name

Filters
Admin in retail tech support Belong to a discussion group Last time I looked up product information, I looked at certain documents and types I dont want legal information emphasized and Im not an expert on this product

32

Summary
 Successful Content Management requires a taxonomic dimension  CM companies have recognized this and added features  Next Step: Content management as infrastructure platform  Need: well articulated intellectual infrastructure  3 important terms: Contexts, Taxonomies, Intellectual Infrastructure  Your choice go back to file management or forward to infrastructure content management

33

Questions?
Tom Reamy [email protected] KAPS Group Knowledge Architecture Professional Services http://www.kapsgroup.com