SharePoint 2010 Findability
- 1. Metadata and Findability in SharePoint 2010Dave MaskellEuropean Sales Managerdave.maskell@smartlogic.com
- 2. Smartlogic SemaphoreSemaphore improves Findability in SharePoint and other applications, through automated content classification and search enhancement.Findability: brings people and knowledge together. Enables users to find the information they seek and promotes content to the relevant audience.
- 3. Customer Benefit ExamplesInternal Knowledge ManagementSharePointInform/Serve CustomersMonetise Information
- 4. Knowledge Management TrendsIncreasing demand for “information about information”Support knowledge workers
- 6. Legal (e.g. eDiscovery, Freedom Of Information, Audit)Systems consolidationFrom departmental to enterprise repositoriesLittle editorial controlPeople are encouraged/compelled to upload everything
- 7. No SEO on the intranetGoalKnowledgeInformationHiddenDataUnknownKnowing What You haveThe British Museum has a collection of 7 million items, of which 50,000 are on displayUS museums have 146 million items, 53% of which are not cataloguedSurfacedKnowledgeYou know where it isValueInformationYou know it’s somewhereDataYou don’t know it exists
- 8. Free-Text SearchThe “long tail” problem20% of content “wins” 80% of searchesFree-Text search is a blunt instrumentHeavy documents float, light ones sink!What are the “right questions”?
- 9. Language and MeaningAboutness- what is it?Terminology - what do we call it?Netbook
- 10. Laptop
- 11. Notebook
- 12. Apple
- 13. MacBookAmbiguity – what did you mean?
- 14. Precision and RecallWhen is a Lift not a Lift?Poor Precision – “false positives”Document mentions Lift once in 100 pages – it is not about LiftsWrong kind of lift – document about “valve lift” systems for carsPoor Recall – “hidden knowledge”Oops!
- 15. Investing in ContextBroader TopicsRelated TopicsNarrowerTopicsTopic-RelevantContent
- 17. The Value of MetadataStandardisation and ContextStandardise Naming & Context:
- 18. Preferred Term = Elevator
- 19. Synonym = LiftTaxonomy/OntologyTerm SetModel MetadataTag Content for “aboutness”
- 20. Based on weight of evidence
- 21. Some documents are about Elevators others are about Valve-lift SystemsClassificationEnterprise KeywordsItem MetadataSearchEnhancementImprove Search Recall:
- 22. “Did you mean elevator”?
- 25. Improve Search Recall & Precision:
- 28. SharePoint Metadata 1.0Structural attributes: Type, Author, Date etc.
- 29. SharePoint Metadata 1.0“Aboutness” (Term Store & Keywords)
- 30. SharePoint Metadata 1.0Column Filters
- 31. SharePoint Metadata 1.0Free-text search, topic refinement
- 32. Metadata 1.0 Challenges“…many departments and organisations in the public sector arbitrarily pick a convenient high-level term from IPSV* to classify all their web pages just so that they can tick the box…” Human tagging is costly, unenforceable, inconsistent
- 33. Who tags legacy/imported documents?
- 34. Changed documents do not get re-tagged
- 35. Term Set changes make metadata obsolete
- 36. Untagged or badly-tagged documents are “filtered out” or waste user time/attention*IPSV = UK Integrated Public Sector Vocabulary (3,000 terms)
- 37. Semaphore forSharePoint 2010Complete, consistent, correct MetadataImproved Search Experience
- 38. Complete, consistent, correct MetadataClassification RequirementsAutomatic Classification is essential for complete, consistent and correct Metadata.
- 39. Automatic Classification needs to include a number of capabilities:
- 40. Quality Classification, not just keyword matching
- 41. Classification triggered on document creation and update
- 42. Ability for users to add/delete tags
- 43. Protect manually tagged documents from reclassification
- 44. Bulk Classification of librariesComplete, consistent, correct MetadataModel-Driven ClassificationHow would YOU decide whether a document is ABOUT Apollo 11?Encapsulate this knowledge map in a formal rule
- 45. Complete, consistent, correct MetadataModel-Driven Classification
- 46. Complete, consistent, correct MetadataUser Classification
- 47. Complete, consistent, correct MetadataProtecting User Choices
- 48. Improved Search Experience“Did you mean?”
- 49. Improved Search ExperienceSemantic EnhancementMetadata SearchRelated TopicsRecommended Links
- 50. Improved Search ExperienceRelated Topics – Guided Discovery
- 51. Solution ArchitectureSharePoint 2010 and SharePoint 2007Smartlogic SemaphoreContentClassificationTerm StoreManagementSearchEnhancementOther Enterprise Applications?
- 52. Enterprise-Wide Hunger for Context…EnterpriseSearchContentManagementPortalInfrastructureDocument Management????SharePoint?RecordsManagement?PublishingSystems?DigitalAssetManagement???Process Management &WorkfloweDiscovery
- 53. …requiring an Enterprise SolutionEnterpriseSearchContentManagementPortalInfrastructureDocument ManagementSubjectGeographyProcessMetadata(Tags)Model(Ontology)SharePointRecordsManagementPublishingSystemsDigitalAssetManagementSemaphore EnterpriseSemanticPlatformProcess Management &WorkfloweDiscovery
- 54. Case Study:UK National Health Service
- 55. Medical Information for PatientsSharePoint with GSA SearchTopic FiltersRecommendation
- 56. Interactive Term Map - Visualisation
- 57. Interactive Term Map - Visualisation
- 58. Social Media Content Classification
- 59. Medical Information for PractitionersFAST ESP across 100+ web sites
- 60. ConclusionsThe SharePoint 2010 Term Store and Managed Metadata are a first step towards improved Findability. To be effective, Metadata must be complete, consistent and correct. This requires automation.SharePoint 2010 findability can be further improved through semantic concepts like “did you mean?”, metadata searching and related topic suggestions.The same is true for other applications that use unstructured content.Smartlogic Semaphore provides an enterprise platform for metadata management and semantic search enhancement.
- 61. Thank YouDave MaskellEuropean Sales Managerdave.maskell@smartlogic.com