To folder or not to folder ... that is the question

Oh, did I mention how metadata turbo-powers:

• Data Loss Prevention
• Records management
• eDiscovery

Quick case study. We created a project management solution for a client. The doc libs:

• 3 folders: project management & control; work files; and deliverables
• project managers place documents in one of 3. These are natural high-level classifications for project managers. If they do it, populating metadata is touchless. Governance!
• All metadata is based on term sets (don't have to be, could be site or list columns-[strongly recommend root site columns if list columns]). Column Default Value Settings automatically applied to the files
• Resulting in every file assigned following metadata
  • Client name
  • Project name
  • Project type
  • Asset classification (e.g., Deliverable, Project Management & Control (like Statements of Work & project timelines)
• On teardown files tagged deliverable and management and control moved to a Records Center serving as a knowledge base. No folders! Hold on to the work files for 90 days, then delete. 

BENEFITS

• Vastly simplifies the tracking and QA work of the Project Management Office (PMO)
• Automated teardown using Information Management Policy and workflows
• Putting files in the 'right' folder often deeply ingrained end user behavior since the early days of on-prem file shares. It's natural
• Discoverability! New PM, assigned a WordPress project, bellys up to the knowledge base searching for prior art. Hundreds of WordPress projects in the KB.  PM specifically seeking examples of user manuals
  • Scenario 1 - no metadata, just keyword search
    
    • usual oceans of documents and file names with WordPress returned. Hundreds and hundreds. User tries to narrow down: "WordPress Manual" - maybe - if authors used exact phrase somewhere in their files. "WordPress User Manual". Maybe proximity searches if power user
    • refiners of some help: date ranges, document types (Word files, PDFs)
    • difficult finding the right set of related artifacts that can really speed the plow - such as Lessons Learned
    • too much work poking and prodding? Abandon ship. Go it alone
  • Scenario 2 - Metadata + keywords. The refiners & keywords
    • Project types = website & WordPress (because of so many WordPress projects, it's added as a website sub-term in the term set)
    • Artifact type = deliverable
    • Enterprise Keywords (aka folksonomy)
    • Keywords & phrases = "User Manual" "Manual" "How to"
    • plus standard refiners like doc type (Word, Excel) & date ranges

Which scenario surfaces the right content fluidly, almost instantaneously? Is likely to see more user adoption by PMs and the PMO? 🙂