Previously, we looked at best practices for Search as well as Folder/Document viewing in a Content Services application.  However, adding content to the system is equally important.  After all, if users do not have an easy-to-use interface that properly indexes and stores content into the system, the search and folder view interfaces will be incomplete due to missing content.  Therefore, we see two main goals for the Add Document interface:

1. Provide an easy-to-use interface that promotes adding relevant content to the system.
2. Index the content properly into the system to make it easy to find the content later, which typically means assigning the content the correct type and metadata.

Or, in other words:

1. Ensure content gets into the system.
2. Ensure content doesn’t get lost.

In this post we will overview best practices that promote the goals above.  Note that in this post we are mainly focused on the user interface for a Content Services application.  While we will not be discussing migration processes or automated ingestion feeds, back end code for these services should still strive to meet the primary goals outlined above.  If you are looking to migrate to a modern Content Services platform, be sure to check out our Top 10 ECM Migration Tips for Success post.

With the Add Document UI goals in mind, let’s dive into the best practices.

Best Practice 1: Index Document During Import

If content can’t be found, it may as well not exist.  Therefore, our top tip is to make sure content indexing happens in tandem with content import.  Content indexing typically includes:

• Setting the correct object type.  Is the document a Vendor Contract?  A Claim Document?  A Standard Operating Procedure?
• Setting valuable metadata.  We recommend requiring as many metadata fields as possible since optional fields are typically skipped resulting in incomplete data.  A proper object taxonomy balances required and optional fields to provide information for search while minimizing friction when importing.
• Placing the content in the correct repository location.  To maintain optimal repository performance, it is recommended to enforce proper content location within the folder structure.

One mistake that we often see is that the UI will simply allow users to drag and drop a folder or document into the repository as the only step in the process.  While this certainly satisfies the goal of making import easy, there are a number of concerning limitations:

• The document type is never specified.  Sometimes this can be automated with code, but often we see customers simply leave the type set to the base “Document” type, missing out on additional rich and relevant metadata.
• Metadata is never specified since the folder structure is describing the document categorization (department, year, customer ID, etc.).  While the user could edit document properties after the initial upload, relying on this second step is not recommended.  The chances of this step happening are low, and if it doesn’t, metadata is limited to the file name. Additionally, misfiled documents have a very low probability of being located.

With the type and metadata likely not set properly, Search loses much of its value to the overall system.  To combat this situation, we recommend requiring proper content indexing at the same time as the import in the Add Content UI.  Since, admittedly, this step does add a bit of friction to the import process, most of the other best practices we recommend concern Goal #1 – making it easy to add content.

Best Practice 2: Display the Document During Import

Even in some systems that follow the above best practice to enforce indexing during import, we often see a big mistake made: after the file upload, the user is taken to a new screen to input the metadata, but the document is not visible.  Instead, we recommend that the document be immediately visible to the user so that he or she can see the document contents at the same time as the metadata indexing form.

In the above example, the user has full context in a single view, allowing them to validate and populate metadata easily.  This UI construct is also an example of Multi-Perspective viewing that we discussed in the Folder and Document View Best Practices post.

Best Practice 3: Leverage Context to Limit User Decisions

The easiest way to limit the friction when adding content to the system while also properly indexing, is to limit the number of decisions that the user needs to make.  This step can be done by leveraging the application context to make these decisions for the user.

First, the system should ensure that the object type is set properly.  The easiest way to ensure proper set up is to automatically select the type based on the user’s context.  For example, when adding a document to a Claim Folder in the Alfresco Content Accelerator (ACA) Claims Management accelerator, the user doesn’t need to choose the new document type at all.  The system knows that since the user is adding a document to a Claim, then the type of the document should be ‘Claim Document’.  While this may seem trivial and intuitive, we have seen many instances where context is ignored, and users are presented with a list of over 50 types to choose from.

Second, the system should make it easy to set required metadata fields.  Once the type is known, the system should smartly display a form for the user to input or verify metadata.  The form should make it clear which fields are required, and should leverage dropdown lists, date controls, etc. appropriately.

Best Practice 4: Automatically Fill out Metadata Based on Context

Continuing from the previous point – displaying a context-aware form to capture metadata, the system should automatically fill out this metadata as much as possible.  This practice reduces the number of clicks and user decisions, which also reduces errors in indexing.  There are a few ways that metadata population can be automated.

One approach that ACA uses is to match common metadata on the document type to the context folder type.  For example, in the Claims Management accelerator when adding a claim document, the system smartly knows that many of the document fields should simply be inherited from the containing Claim folder.

The fields in the red boxes above have been automatically populated from the parent Claim folder.  Based on system configuration, some fields are read-only, whereas others can be modified during import.  Whether an inherited metadata field is editable or not can be configured based on the system needs, balancing data integrity with flexibility during import.

Another approach is to smartly fill out metadata via code.  For example, leveraging OCR along with AI/ML to automatically pull content from the document into properties.  For example, the code could smartly pull off vendor information, invoice number, amounts, etc. from an invoice document.  

ACA also has an option that allows users to highlight or double click values in the document and the system automatically populates the metadata form with the text in the document. Allowing the user to click to pull data directly from the document content is another way that small UI features can simplify the import process and have a big impact on user experience and acceptance.

Best Practice 5: Automatically File Content Based on Metadata

One of our Folder and Document Viewing best practices was to leverage Search instead of relying on a folder tree.  In fact, we recommend that a content services UI does not provide a full folder tree viewer at all.  However, it is important to keep the underlying folder structure clean and efficient for repository performance.  The best approach for filing documents in the repository is to use an ‘autofile’ tool to automatically place folders and documents into the correct locations, rather than relying on the user.

In a case management scenario, the user can upload documents directly to the case folder.  For example, a Claim folder may be automatically created in the repository based on metadata.  Typically, this could be based on the claim number and/or loss date.  In this scenario, any documents are uploaded to the claim folder location automatically without the user needing to know where the claim folder actually resides within the repository structure.

In document-centric scenarios, we recommend that the autofile service base the final location based on metadata.  For example, in ACA’s Policy and Procedure Management accelerator, there’s an “Add Document” action in the header.  When a user uploads a document, the autofile code takes the document’s Category, Region, Site, and Business Group property values and automatically places the document into the correct folder.  If a document is added for a new Site, for example, then the appropriate folder(s) are automatically created in the repository.

Auto-filing content eliminates the folder navigation step, where users are often overwhelmed with options, while also ensuring that the correct metadata is applied to the content so it can be easily located in the future.

Best Practice 6: Smartly Ingest Content to Include Related Information and Create Context

In both our Search and Folder/Document View best practices, we mention taking full advantage of repository relationships to automatically surface information to the user based on their context.  While relationships can be managed either via code or manual clicks after a document is imported, we recommend smartly ingesting content to automatically create relationships.  Ingesting content and relationships simultaneously creates context for related information immediately, which can then be leveraged throughout the application.

One example is ingesting emails. There are a number of automated steps we can take during the import process to create relationships.  Attachments can be automatically pulled off from the email and stored in the content repository.  A parent-child relationship is created from the email to the attachment(s).  With this information, the UI can easily display an attachment list when viewing the email. Attachments can also display the inverse relationship, linking them back to the email.   With emails in particular, It is important to be smart about what constitutes an attachment.  Certain image attachments should be ignored such as corporate logos, social icons, or other ‘signature section’ elements.

Full text search can be leveraged as well.  In the SOP use case, when uploading a procedure document, an automated full text search could scan the document to look for related SOPs and Forms.  These relationships can be automated based on the result of the search, saving the user time and effort in creating the links between documents.

Summary

One of the key pillars of any content services solution is enabling users to easily and accurately locate content. By nature, adding content to a repository will involve some extra time and steps to classify and tag content.  These steps, automating them where possible, is important to ensure that content is properly indexed and easy to find.  This post points out many small but important ways you can simplify the process of importing content so your end users will embrace the application and ensure the original goals of adding and organizing content are met.

Remember to apply these best practices to all content entry points into the system to ensure that the Search and Folder/Document view interfaces perform accurately and consistently.

Comment your thoughts below or on LinkedIn. Do you have any folder and document viewing best practices to share? Let Docuvela know!