Using Semaphore for Microsoft SharePoint

Reduce The Burden of Tagging Content with Semaphore for SharePoint

In any corporate SharePoint implementation, metadata is critical.  If you know what something is about, you can put processes in place to manage it and make it more findable.  SharePoint has the capability to develop workflows and search content, but relies on users to add the metadata.  Semaphore, the Enterprise Semantic Platform, from Smartlogic is a complete, comprehensive solution to address the metadata gap in SharePoint 2007 or 2010 and provides:

• The highest quality of automatic subject classification available.  Semaphore’s unique rules based classification is accurate, fast, intuitive, controllable, and yet requires very little setup or training time
• Comprehensive text analytics including entity extraction and natural language processing tools to deliver better performing taxonomies in a fraction of the tim
• Full lifecycle “develop, enhance and maintain” taxonomy management and governance integrated to SharePoint preventing the chaos caused when anyone can modify shared taxonomies
• Components that semantically enhance the user experience in any consuming application to deliver the most powerful and flexible navigation along with the most accurate search results.

File Plan versus Metadata

If you’ve tried to organize your content in SharePoint by sites and a file plan you’ll have realized just how restrictive that can be for the users.  One user’s filing system is a mystery to another, so the goal of improving re-use, collaboration and knowledge share looks distant. Metadata is the answer and Semaphore allows users to “tag” their content and access it via the more dynamic views available in search and aggregation web parts.

Subject Taxonomies versus Classification

The most useful metadata element for find-ability is a subject based view. Subjects are typically organized in a taxonomy– a hierarchical grouping of the concepts within which the user can drill in to find the concepts that match their document. However there is a problem: make the taxonomy sufficiently high level that it is easily understood by the users and you end up with huge numbers of documents tagged under each node; extend the taxonomy to a suitable level of granularity for the content and the user can be overwhelmed (especially using SharePoint’s very limited out-of-the-box hierarchical list components). Semaphore allows for poly-hierarchy, meaning a term can have more than one parent e.g. Arnold Schwarzenegger could be a narrower term of Actors, and of Governors. Poly-hierarchies provide greater flexibility in modelling taxonomies so that a single term can be navigated through multiple paths to make a user’s journey of the model is more intuitive. In addition, poly-hierarchy ensures that there are no missing results.

Reducing the User Burden with Semaphore for SharePoint

Provide the user with a familiar tree-view taxonomy browser, or even better, searchable taxonomy components when they add a document and the quality of the tags added will improve. Better still is to automate the process - use natural language processing and a rules based classification process to suggest tags automatically and the burden on the user is completely removed.  Automatic classification might be the only option to add the subject metadata against external content sources (where no one internally is sufficiently familiar with the content to tag it) or on large volumes of legacy data (where the cost of a person opening, reading and tagging thousands of documents is prohibitive).  SharePoint 2010 includes the Content Organizer: the system designers create rules that determine where in the file system a document should go based on its metadata.  Semaphore can automatically classify the content on upload and thus adds intelligence to this routing feature.

Consistent Metadata

SharePoint 2007’s design ensures consistency of metadata within a Site Collection.  In our experience organizations are deploying SharePoint using tens or hundreds of site collects.  A seemingly simple task like ensuring the items on a list are updated consistently in every library they occur is a challenge using the basic tools.  Abstracting the management of metadata to a central service (like Semaphore) resolves this issue.

SharePoint 2010 has a central service called the Term Store, but its implementation can again raise questions of governance.  It is possible for administrators to create their own term sets from the store.  While the terms will be the same, they will be allocated separate term store IDs.  In one early implementation we’ve heard of, the term “Business Intelligence” had been replicated 19 times.  So a user subscribing to an RSS feed expecting all information on that subject would only receive a fraction of the possible content.  Semaphore resolves this issue by providing stricter governance of a corporate taxonomy.

Precise and Complete Results

Two goals organizations aim for when using taxonomies and meta-data tags are to increase precision and to increase recall (completeness).

Precision Tagging content ensures that results don’t include content that contains the search keyword but in the wrong context. For example, in a search for “Orange” (the telecommunications company) should not include content about the fruit or color.

Recall Tagging content ensures that results include content that doesn’t contain a search keyword but is clearly about the topic the user seeks. For example, a search for “garbage” should include content that uses terms such as trash, litter, recycling or waste in the correct context.