Best Practices for ECM in SharePoint and OneDrive for Business

As we announced today on Office Blogs, we are happy to see SharePoint evaluated as a Leader in The Forrester Wave™: Enterprise Content Management—Business Content Services, Q2 2017.

Our pivot to content services began at Ignite 2016 in Atlanta where Jeremy Mazner and I presented a breakout session on SharePoint ECM. You can watch the entire session below:

Video provider requires cookies to play the video. Accept to continue or watch it directly on the provider's site.

Accept

At Ignite 2016, we reviewed some best practices for traditional ECM and Content Services.  Let's take a look.

Minimize ECM in default libraries.  The default team site document library is the common repository for many other collaborative elements of Office 365.  Creating a team site automatically enables other apps, such as Planner and Teams, that sit on top of the SharePoint foundation; they all use the default document library for the team site.

What does this mean?  For starters, avoid heavy customization of that library.  Adding a lot of custom metadata columns and making them required will have unanticipated impacts on those other apps.  If you want to create custom ECM processes and data, just add a new library to your site, or use a repository template like a Document Center.

Reuse common fields as site columns.  We demoed this at Ignite as well.  Although there’s nothing wrong with using a one-off column on a single document library, many times the same field is needed in more than one place.  It’s a best practice to create these fields as site columns so they can be shared through the site collection.  And if they require “choices”, defining them as Managed Metadata columns is the preferred way to assume a single definition can be shared through the information hierarchy.  With the new release of Feature Pack 1 for SharePoint 2016, you can also share those metadata definitions with an on premises environment.  That means you can add the same look up value to all your content in libraries on premises and in the cloud.

Do not overwrite default content types.   95% or more of all content types used in SharePoint libraries are the default content type, “Document”.  It has no template, no required metadata, and no retention policy by design.  Although you may be tempted to updated this content type, resist the urge!   Instead, create a child content type that inherits its base definition from Document.  That way, Microsoft updates can’t override your customizations.  And if you put the content type in a Content Type Hub, you can share that definition through your on-premises or cloud architecture.

Minimize MMS complex pinning, reuse if hybrid is on roadmap.  If you’re planning to rollout Hybrid Taxonomy, the simple design pattern is that the cloud is the “parent”, and on premises term sets are the child.  That means local techniques for reusing terms inside the same architecture, such as pinning, create a local “master” tag which will be overridden by the online master.  We recommend using the PowerShell command Copy-SPTaxonomyGroups to copy preexisting terms from on premises to SharePoint Online as the new “master”. 

As a result, try to unwind any pinning or reuse until you’ve created a hybrid architecture first.

Strategize use of hub-based content types (don’t overdo it!)  We remind you about stories about the companies with, for instance, 1400 content types because they are great cautionary tales.  Even if its supported, conceptually it’s hard for users to leap from “Document” to 1400 or more syndicated content types.  Ideally, you syndicate the most heavily and widely used types of content (contracts, presentations, policies) while allowing for localized definitions or inheritance to accommodate regional or divisional variations.

Hashtags, keywords?  Be careful…  Ad hoc classifications such as Delve boards, social hashtags or Enterprise Keywords let users create any possible term or variation including misspellings.  Providing content services at scale across the enterprise usually mandates a higher level of precision and predictability.  The Managed Metadata Service (MMS) is likely a far stronger framework for classification in ECM scenarios.