Is a tag based system a better UX than traditional folder structure?

Question

Background

We are a manufacturing company that deal with mainly aerospace, automotive, and assembly type customers. Our users have been using a network folder structure for as long as I have been here (over 15 years). It's messy, hard to find things, and very deep in structure. The structure comes in the form of:

Project A
  -Folder
     -Folder
        -Folder
          Files
       Files
    Files
   -Folder
     Files
   -Folder
  ...
Project B
  ...
Project C
  ...

Problem

• In around 60% of cases, we run through the limit of file path/file name sizes of more than 255 characters. Folder names are like sentences in some cases such as Automotive Company X - Transmission assembly V8 line
• We are constantly running out of network space
• It is hard for anyone to find anything
• Files can easily be moved around by "accident"
• The depth of the folder structure is so long that sometimes you forget where you are and other times you end up in a folder that has nothing!

Proposed Solution

An internal CRM type system (think Salesforce or SugarCRM) and within this CRM users can manage their projects. This app uses various search technologies so when you need to find a project you can easily find it with just a few keystrokes.

Default view

Users go to the project in this application and can add documents:

Upload

Instead of a folder hierarchy, users upload [multiple] files and requires them to "tag" (similar to how we tag questions on here):

View and Filter

In the above example someone entered the tag "issues" and found all documents where this tag appeared. In this case one file is returned.

The goal is to eliminate the confusing folder structure, and make it easier and quicker for users to find existing projects, and upload existing projects.

Question

Assuming it resolves the current UX problems we have, can anyone see any new UX problems it could cause?

What are the benefits of retaining the tried and tested folder structure?

Answer 1

[...] my goal was to eliminate the network shared structure.

It'll help to center the goal around a clear pain point. You mentioned issues of 'g-drive' but the top two issues (length of the file path and storage space) are technical and not related to usability. From your description I think it's getting to the right information quickly that's the biggest issue. You're required to go through a barbarian sequence of steps before you see any results.

So the focus here is to improve navigation / exploration. I think your approach has potential to improve on this aspect. First few things that I think are working well for you:

• You're flattening the hierarchy: it reduces the number of required steps.
• You show results on a project level, and possibly tag level: you can decide at any point to stop going deeper and scroll through the current results.
• You require users to assign tags which helps organize information space.
• Separate pages for projects. This helps frame and contextualise further navigation.

To make navigation mentally and physically less demanding I'd make the following recommendations:

• Filtering the way you demonstrated above requires user to build exact specification of where the file is. That's hard. I'd recommend to enable active exploration: show user available tags, file types, creators, date selector etc. This way the user can quickly check different locations of your information space.
• Make the above filters present at all times, starting with a project page. Consider using multi-faceted navigation.
• Show exploration history. You could use breadcrumbs for example. Make them clickable (and removable) so that users can quickly go back.
• If documents can have multiple tags assigned show which tags co-occur with the selected tag. This way you'll help users narrow results in a meaningful way (never leading them to 0 results).
• Implement search suggestions. Suggest documents (possibly grouped by tag) and tags themselves. Make full-text search available. All this will let users bypass sequential exploration if they have sufficient detail on what they're looking for.
• When users assign tags to documents show the available tags. Don't hide them in a drop-down. Bear in mind that tagging is not a silver bullet to the problem of document organisation. People often use different tags to express the same thing. You might want to make the creation of new tags more explicit - i.e., introduce a bit of process here.
• Consider reusing tags across different projects if there's a big overlap.

One of the major risks is the simplicity of file operations in 'g-drive' vs web browser. Moving files on the desktop, creating new folders, deleting them etc. is extremely easy and natural. You'll have to make sure that the corresponding operations in your web application are equally easy.