I had a conversation with a colleague recently about the use of content types and site columns for cataloging content (tagging) in SharePoint. We both agreed that cataloging content improved the search relevance of content over not cataloging at all. The point of disagreement was whether or not it is better to use site columns exclusively or to use a content type that is inherited from the Document content type and is a more specific content type to the data that you are cataloging. If you have been reading my blog for more than a day, you will understand what my argument was.
The basic question is whether or not cataloging content with a site column is equally or more relevant than cataloging the same content with a content type when all other factors are equal. With this in mind, I created an experiment in my Virtual Machine (VM) to determine which method yields the better search results.
Description of Content
For this experiment, I utilized the content of the Grouped d20 System Reference Document; specifically the Basic Rules. For those who are not aware, these documents are the open source rules for the Dungeons & Dragons role-playing game.
Setup
To ensure that all other things are equal, I started by creating a brand new Web Application. For the site collection, I selected the Publishing -> Collaboration Portal. This portal provides a Document Center where I can create document libraries to hold my documents. Given that click distance and other architectural factors affect relevance, this ensures that I can create document libraries that will be equal in these considerations.
Create the Document Libraries
After creating the application to hold my experiment, I opened the site and created two document libraries in the Document Center: once called CT_Documents to hold documents cataloged with content types and one called SC_Documents to hold documents cataloged with Site Columns. Next I uploaded all of the documents from the Basic SRD into both libraries.
Create the Metadata
In this step, I am creating a site column to describe the type of rule set these documents belong to and a content type to do the same. I created a choice site column called ‘Rule Set.’ The choice list included all of the groups of rules that are defined on the d20 site: Basic Rules, Divine, Epic, etc..
Next I created a content type for Basic Rules documents. I inherited this content type from the Document content type and named it ‘Basic Rules.’ For the purpose of this experiment, I made no other changes to the content type and added no additional columns. The result is a content type that looks exactly like the document type but has a different name.
Catalog the Content
After creating the metadata types for the experiment, I went into each library and cataloged the content. In the SC_Documents library, I added the Rule Set site column to the Document Library using the library settings. This allowed me to assign a Rule Set value to each of the documents in the library. I did this by selecting Edit in Datasheet view in the Actions menu; then I set the Rule Set property of each document to ‘Basic Rules.’
In the CT_Documents library, I added the Basic Rules content type to the Document Library using the library settings. This allows me to set the content type of each document in the library to Basic Rules. I then edited the properties of each document and selected the Basic Rules content type. At this point, I have not entered a title for any of the uploaded documents.
Crawl the Content
It is important to understand that this is an out of the box search of the content. I have not defined any authoritative pages, setup crawl rules, or done anything that may impact the final relevancy. With this in mind, I ran a full crawl of the ‘Local Office SharePoint Server Sites.’
Conduct Searches
Finally, I ran searches in Search Center using the document names as the search terms. In every case, the Content Type always lists ahead of the Site Column. The results tend to support my argument that when you are searching for a type of document in your system (e.g., project plan, design documents, requirements) it is better to catalog those documents with a content type than to use a site column as a cataloging element. You will ultimately improve your search relevancy for those kinds of documents.
Endnote:
To be clear, my colleague had a very good point regarding the content type versus site column debate. "Don’t get carried away defining content types." Her concern with my approach is that I would be defining a content type for each document that I uploaded into my SharePoint solution. This is certainly a risk but one that a content designer must be aware of. As I have stated in earlier posts, the content type should identify a category of content; it is the job of the site columns to identify a specific piece of content.
