Tagging in the corporate environment


Right, so if you’ve read Tagging 101 and The Glory of the Garden we should all be on the same page now? That or you’ve flamed me and gone somewhere else. If you’re still reading I assume we are in general agreement so far.

What is special about the corporate environment?

Tagging heterogeneous assets

Flickr, Gmail and del.icio.us all allow tagging of a particular object: photographs, emails and internet links respectively. Corporate networks contain assets of many kinds. Those listed are joined by documents on fileshares and objects in databases and applications. Should we implement a separate tagging mechanism for each asset class?

Documents

I mentioned in Tagging 101 that there was no good way of tagging your local documents at the moment. We are stuck with the file name and the folder hierarchy they live in to classify them. This is bad on a personal scale, but in a corporate environment with hundreds of teams working on thousands of document sets it is a ridiculous way to manage documents.

This is why document management tools gained some traction a few years ago. What they provided was, seemingly, exactly what we are after – a way of appending multiple classifications to a document so as to be able to extract any subset with simple filters. The way they did it was to seize the document, put it in a walled garden and restrict the ways you could access it to conform to their modus operandi. There are some good reasons for this: if you implement a tagging system separate from the document storage then you are open to the link becoming stale if a document is moved.

The operating system vendors are aware of this. Microsoft has repeatedly tried to add this functionality to their filing system but has never got WinFS actually released, either separately or bundled with Windows. Apple are working on a similar concept that may be in OS X 10.5 or more likely OS XI (or whatever). No doubt Linux file system coders are catching up fast.

The key thing is to know why you are tagging a document: is it so you can be sure of finding it again later, or is it so it comes high up in a general search for relevant material? For the former you want a sure-fire way of knowing that whenever you enter that tag, the document will appear. For the latter it is probably OK if the odd document drops down the search results due to losing its tags.

Objects in applications

A portfolio of applications in any corporate environment is likely to have several key systems whose data is consumed by many other applications. Many corporations invest large sums of money in enterprise data management solutions to ensure that the key objects like customer and contact data, product data and transaction data are warehoused centrally and consumed widely.

It would surely be beneficial if there was a mechanism whereby a user could tag these objects with keywords that were meaningful to him or her. Some functions of a CRM system could be replaced by allowing people to, say, make their own christmas card list by tagging contacts with “Xmas card” (or “holiday greetings” if they are PC enough). Maybe not a great example, but one can see the principle and the opportunities such a general mechanism would present.

I do not believe there exists a tagging system that will allow us to aggregate the tags from all these sources, still less one that allows us to centrally manage the tags on heterogeneous assets.

Respecting access control lists (ACLs)

This is a challenge for any public paradigm that wants to live in the corporate ecosystem. Public search engines tag what is public. It is a simple binary decision: can I see it? Yes, so index it.

A corporate search engine needs privileged access to everything, but needs to restrict who sees which search results. It needs to preserve the ACL of each link it discovers and match that with the credentials of the person doing the search.

The parallel with corporate tagging is exact.

You should be able to tag what you can see (whatever it is) and manage your tags like any set of bookmarks. However much of the value of tagging for the enterprise lies in the aggregation of those tags over the whole population of users and this cannot be done without preserving the ACLs of the assets being tagged. If I wanted to see what other people had tagged with “merger rumour” then the results have to depend on my access to the tagged documents, emails, whatever.

Integration with other tagspaces

If we cannot create a single uber-tagging system for the entire enterprise then the task remains to integrate all the separate tags from individual systems. This can be achieved with the right enterprise search platform, which can be trained to understand a search term such as “tag:merger_rumour” as relating to intranet page tags, SocialText page categories, Sharepoint document categories, CRM system tags and others, and integrate them in the search results.

We should also recognise that the external tag-based systems may also have value in this context. The analogy here is with the internal knowledge-base wiki and the external Wikipedia. If there is a perfectly good article in the Wikipedia then there is no need to write one for the internal wiki, and if it doesn’t exist then write one. Externally. If there is specific local knowledge of commercial sensitivity then a codicil can be written for the internal wiki. It is only for purely internal knowledge that a full internal wiki article needs to be written. The challenge is to combine these two resources seamlessly into a single knowledge base.

Again there is an exact analogy with tagging systems. To tag an internet link with a general-purpose tag might as well be done in del.icio.us, then it is available to you even outside the work context. Tags with an internal meaning should be done in the internal system even if it relates to an external resource.

But wait – this is too complicated for people to operate! In the end people will tag wherever it is easiest for them to do so. We have to accept that and make sure that a search produces useful results wherever it finds the tags. If we enable heuristic tools like tag clusters and tag clouds in the corporate environment they need (optionally to the user) to include tags from external systems too.

It’s a sizeable challenge. I don’t see any of the current contenders (dogear [more here] or Cogenz) really stepping up to the plate. It’s taken a long time for the search engine people to understand the corporate space, I think it will take an equally long time for vendors to get to grips with enterprise tagging to deliver its full value.

Advertisements

0 Responses to “Tagging in the corporate environment”



  1. Leave a Comment

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s





%d bloggers like this: