Case Study: Oracles Knowledge Based Support System

1
Case Study Oracles Knowledge Based Support
System

• Brian Bombard
• Director, Global Knowledge Management
• Oracle Corporation

2
Agenda

• Pre-2000 Knowledge Management
• Oracle Support Services Evolution
• Knowledge Management Today
• Knowledge Management Learnings

3
Oracle KM Program Pre-2000

• Product Line Specific KM Programs
• Dedicated team of publishers
• Processes similar across product lines
• High Re-Discovery Rates
• Content not available for customer use
• Inconsistent Capture of Knowledge
• Regional/Center Centric
• Unclear accountability for the knowledge base
• Organic knowledge creation and maintenance

4
Oracle Support Services Evolution
Web Service
Call Center

• 95 New SRs Online
• 113M web knowledge Interactions

• 2M New Calls (SRs)
• ?5 Online SRs

5
KM Challenges

• Accountability
• Needed a global ownership model in place to have
  accountability of knowledge areas
• Scalable Infrastructure
• Infrastructure to support and optimize knowledge
  management activities
• Systematic way to identify opportunities
• Process and Standards
• Uniform and consistent KM processes across
  product lines and regions

6
KM Transformations

• KM Community Establishment
• Global ownership and accountability for a product
  area
• Each member of the community dedicated 4-8 hours
  a week
• Root Cause/Corrective Action Process
• Process for the historical analysis of Service
  Requests to determine knowledge and/or product
  gaps
• Corrective action creation and execution
• Knowledge Improvement, Product Enhancements, etc.
  
• Centralized KM Program Team
• Centralized team created to bring consistency and
  leverage throughout the knowledge management
  program
• Provider of the process, training, tooling, and
  infrastructure to manage/facilitate knowledge
  communities and root cause process

7
Oracle Knowledge Management Today

• Single Knowledge Base For Support Content
• 400,000 Solutions (60 are Customer Visible)
• 113 Million Knowledge Base Interactions/Year
• 58,000,000 MetaLink Reads/Year
• 55,000,000 MetaLink Searches/Year
• 2.8 Million Problems Solved/ Year
• 8.9 of MetaLink Sessions Are Resolved by the
  Knowledge Base
• Customers solve an avg. 3 problems by
  self-service for every 1 SR logged

8
Oracle Knowledge Management Ecosystem
9
Knowledge Management Workflows

• Inside the Resolution Workflow
• Authoring and publishing of problem/solution
  knowledge
• Conducted by all engineers as part of the
  resolution process
• Outside the Resolution Workflow (Proactive)
• Knowledge maintenance and proactive/consultative
  knowledge activities
• Conducted by knowledge community members

10
Authoring and Ticketing Integration
Immediate load from SR into Knowledge Base
11
Resolution Workflow Knowledge Capture
Automatic categorization and filing
Automatic search using real-time signature
generation
12
Automated Referencing and Linking
In-line spelling check and correction
Automatic signature generation and knowledge web
All references are automatically extracted,
validated and added to the final document
13
Rapid Visibility (RaV)

• Transactional knowledge with a relatively short
  life-span
• Important to make customer visible quickly and
  avoid large backlogs
• Certification allows engineers to make their
  articles customer viewable without independent
  technical review or publishing

14
Proactive Workflow
15
Business Need and Requirements

• Alerting Mechanism to Identify What Needs to be
  Done
• Should we be analyzing Service Requests, doing
  feature assessments, processing the backlog, or
  creating new content for new Versions?
• Prioritize Work
• We have limited resources. Which action should
  we do first?
• Proactive Planning and Scheduling
• We plan weeks in advance. What needs to be done
  next month and how much resource is required?
• Return on Investment
• We provided engineers with KM time, what did they
  do with that time? What was the impact?

16
Knowledge Base Business Intelligence

• Opportunity Identification and Prioritization
• Automated business rules run against the
  knowledge base to identify opportunities for
  maintenance
• Opportunities are categorized, classified, and
  prioritized for execution
• Data Clustering
• Application of Oracle Data Mining on Service
  Request signatures to identify clusters of like
  issues for new creation opportunities
• A signature is a data structure to sufficiently
  characterize the object.
• Using a sophisticated pattern matching algorithm
  and a pre-seeded dictionary, all knowledge
  objects (SRs, Notes, BUGs) have a signature
  generated for them.
• Dictionary includes all known Oracle terms and
  common phrases.

17
HealthPlan - Maintenance
18
HealthPlan - Proactive
19
Automated Knowledge Maintenance

• Auto-Archive
• Identifies articles that meet the specified
  criteria and Archives the article.
• An archived article is viewable by customers
  using MetaLink, but is not returned by MetaLink
  basic search.
• A subscription service available to all
  knowledge communities
• Content owners can opt-out selected folders
• Backlog Maintenance
• Automated job that identifies and purges
  abandoned PROBLEM/HOWTO articles in the knowledge
  base
• Author is emailed 14 days prior to any action
  being taken and provided the steps needed to
  avoid purging
• Article is not updated in the last 49 days and
  has not had any reuse internally

20
Summary and Learnings

• Knowledge is a Culture
• Culture is fostered through executive and
  management leadership
• Embed knowledge ownership into the support
  delivery model
• Recognize and award contributors for the value of
  their contribution, not quantity
• Knowledge Base Management Must Be Continuously
  Improved
• Inside The Workflow Creation Needs To Be Easy,
  Integrated, And Streamlined
• Leveraging of tools and automation is necessary
  to optimize process and lower the cost of
  maintaining the knowledge base
• Understand Your Customers Wants and Needs
• Listen, Watch, Analyze
• Leverage customer feedback and the web service
  experience data to continuously improve the
  experience
• Stay Connected

21
Backup Slides
22
Dynamic Article Linking
Error codes and Key Facts are appended to the
document, and become hyperlinks on MetaLink.
Clicking on the hyperlink will retrieve all other
articles linked to the same Key Fact
23
Program Goals and Drivers

• Increase Customer Satisfaction
• Provide current and up to date proactive and
  problem specific content to help customers
  optimize their Oracle investment
• Continuously improve the web interaction
  experience with the knowledge base
• Lower Cost of Ownership
• Reduce the time and resource it takes to manage
  the knowledge base and identify new opportunities
• Prevent customer problems from occurring through
  proactive knowledge delivery
• Showcase Engineer Expertise and Contribution
• Use the knowledge base to showcase the expertise
  of Oracle Engineers
• Recognize and rewards engineers for their KM
  contribution

24
Service Request Signatures

• A signature is a data structure that is assumed
  to sufficiently characterize the object.
• Using a sophisticated pattern matching algorithm
  and a pre-seeded dictionary, all knowledge
  objects (SRs, Notes, BUGs) have a signature
  generated for them.
• Dictionary includes all known Oracle terms and
  common phrases.
• Experts periodically scan content and add or
  delete additional phrases.
• Signatures are created in the background, but can
  be modified by the engineer for optimum precision