Posts Tagged ‘tags’

Organize business collaboration groups using tags

February 13, 2009

Nest folders

One question I get all the time is can we organize groups in nested folders.  We are all so accustomed to this way of thinking about organization, we assume it is the right way to do it.  For example, I want to create my Sales group and have East and West under it , and so on.  Or, someone uses GroupSwim for a project and wants all technical groups under one umbrella.  We can all relate to this habit.  The problem is it doesn’t make sense.  The minute you build a hierarchy for an organization or project, it becomes out-of-date or simply wrong.  There is a better way!

From the beginning, we designed GroupSwim to define relationships between groups through tags, not hierarchy.  The tags determine how groups are related. In the previous example, I could create three groups; Sales, West, and East; each one has the “sales” tag applied to them.  I can also tag the marketing group that works with sales. In this scenario, when I search for “sales”, I get all four groups.  The benefits are legion:

  1. Tags are dynamic, just like business and life
  2. You can have as many as you want
  3. You can change them at any time

By using tags as an organizational metaphor, we allow our collaboration sites and customer communities the opportunity to evolve over time.

The other important factor here is search; it is the great equalizer.  It doesn’t matter where you put stuff if you have a good search engine, which we do.  It also allows the ultimate freedom to organize because in the end, you can find what you need quickly regardless of where it is.

Frankly, the biggest barrier and con to this approach is habit.  People are simply so used to seeing nested folders, that it seems odd to them when they don’t see them.  While we aren’t the only technology using tags versus folders, we are going to stick to our guns on this one.  It does force to do some “evangelism” on this topic, but we can take it.  What do you think?  Miss those nested folders?

Paper on Applying Semantic Technology to Enterprise Collaboration

February 5, 2009

Our very own Jari Koister, GroupSwim Chief Technology Officer and Phd, published a paper on Semantic Universe.  He discusses in detail how GroupSwim utilizes techniques like tag ontologies, natural language process and semantic search to improve Enterprise Collaboration.  Please check it out here and let us know what you think.

Semantic Web Features in Real-life

September 15, 2008

There is a lot of discussion about the semantic web and how it will change the way we use the internet. At Groupswim, we have been considering how to integrate semantics into our feature set. Although we have not promoted it aggressively, we have integrated some features and we have also done some prototyping for future features. That said, we are always concerned about overwhelming the users with incomprehensible features, and we rather try to hide the complexity.

Semantics mean for us that users can associate meaning with items such as discussions, files, wiki pages, tags etc. to improve search and comprehension of information. As an example, GroupSwim currently has a feature called “tag training”. The name is somewhat minimizing the potential of the feature. The feature allows a GroupSwim site to create an ontology of concepts. For example, if you have a site concerned about collaboration and information in the mobile telephony sector, you can relate different terms to each other. “Sony-Ericsson” is a “company”, “Sony-Ericsson” is a “mobile-manufacturer”, Nokia is a “mobile-manufacturer”, “Nokia-N95″ is a “mobile-phone”, “mobile” is a synonym for “mobile-phone”, “Motorola” is a “mobile-manufacturer”, “mobile-produce” is synonym to “mobile-manufacturer”, “mobile-company” is synonym for “mobile-producer” etc.

We realize building a full world encompassing ontology is nearly impossible. So our focus is to enable a site to create ontology for its specific topic area. This ontology then helps the users of the site find information. Let’s say you look for posts about mobile companies. You happen to know about Nokia so you search for that. Obviously you will hit all information tagged with “Nokia”. Thanks to the aforementioned ontology, Groupswim can propose widening the search to “mobile-manufacturers” and “mobile-producers”. GroupSwim can also infer that “Sony-Ericsson” is also a mobile company. So by building up semantic ontologies, GroupSwim can greatly enhance the user’s ability to find related information. Below is a screen from Groupswim showing parts of the ontology we have built up for the GroupSwim internal development collaboration site. It greatly enhances search and discovery by relating otherwise unrelated terms.

Simple Ontology for Groupswim site

Simple Ontology for Groupswim site

These definitions enable search for any information related to competitors. If you search for competition, you will find information about competing companies even if they are not tagged or annotated with the term “competitor” or “competition”. GroupSwim can use the ontology to automatically find the related information.

Going forward, we have plenty of ideas on how to incorporate more semantic knowledge into GroupSTowim. In order to not reveal too much of our trade secrets, I will just give you a flavor of some simple things we are prototyping while also releasing new major functions. One such function is that we can integrate with external semantic databases and allow users to make associations between terms and elements in GroupSwim and semantic definitions outside of GroupSwim. As an example, let’s assume you type in the tag “Volvo”. By retrieving semantic information, we can ask the user to clarify whether they intend the company ‘Volvo Cars” or a specific Volvo car, or perhaps the Latin word “Volvo”. Such an association will clearly help users to search and find information in GroupSwim.

Semantics for Volvo

Semantics for Volvo

Looking forward to hearing how you think semantics can help us improve your GroupSwim experience.

Tag-Based Workflow

June 18, 2008

The thought occurred to me that we could use tags in GroupSwim to manage workflow.  What I mean by this is use tags to dynamically assign work to users/groups/teams in a GroupSwim site and monitor them in real time.  I’m not talking about a structured, pre-defined workflow engine like applications such as a help desk might utilize.  However, using tags and discipline, you could accomplish similar tasks for planning or project work.  I would not recommend this for an operational system like supply chain or customer service, but it could work for less formal activities that may not have the kinds of dedicated systems the aforementioned functions have.  Furthermore, small businesses might not have anything so this kind of system might be better than nothing.

Here is how it could work:

  1. Define how you want the tags to work.  For example, you could use this system for managing issues for a project, for steps in planning a marketing campaign, and other less formal operations.  You should pre-determine status or other kinds of tags that apply to different work streams.  For example, you may want to add Open, Closed, Stalled, or whatever to issues as you post and discuss them so they can be tracked.  You could use a similar arrangement for other activities like NeedCustInfo or whatever.
  2. Have every user/team in the GroupSwim site you use add their Username, first_last name, or some other unique identifier as a tag they watch; they could also add tags for groups like ProjTeam or EmailCampaign3 and every person in this group should add that tag to their list as well.  In my case, I would add the tag JasonR to my watchlist to monitor it.
  3. Each activity should have a manager making sure things get assigned appropriately.  Then, if an issue or activity needs work, the manager or person responsible merely adds the tag of the person who should either work on it and/or monitor it.  As soon as a tag gets applied (JasonR in my case), the person would get an alert through email and/or RSS that there was something they needed to address.
  4. Each time a person finishes a task, they can add the tag for the next person or group that needs to work on it and they would be similarly alerted.  They can also delete the tag for themselves taking it out of their queue.
  5. Once an activity or operation was complete, the manager or designate could add the tag Complete or some other indication that things were resolved.

As you can see, this method would require some discipline.  It’s flexibility is both its strength and its weakness.  I intend to test this internally when the opportunity comes up.  Please let me know what you think and if I’m missing something.

High Quality Tagging Yields Significant Benefits – How and Why We Do It

May 22, 2008

Many of the core features we develop for GroupSwim depend on high quality tags to associate with content.  The tags we produce from the Natural Language Processing get applied to all content added to a group/site.  For example, we tag emails, discussions, files and wiki pages.  The software suggests tags automatically and the user adding the content gets the option of using the tags we suggest, adding their own, or a combination of both.  Our goal is to product quality tagging with this methodology.

The benefits from high quality tagging are significant.

  1. Our semantic search engine finds information quickly and efficiently.  For example, if you search on the word integration, the search engine suggests other words associated with integration allowing a user to narrow their search and find what they want fast.  This would not be possible without consistent tagging.
  2. Topics you care about get tagged every time and are very visible for analysis and response.  You can train the software to key on tags that are most relevant to your group.  For example,  you could add your products or competitors to the tagging engine to ensure the content always gets tagged. Furthermore, you can create relationships with the tags.  For  example, you could train the engine to know “Apple” is a type of “Computer”.  Then, if you search on “Computer”, it will also bring up content tagged “Apple” even if it doesn’t have the word “Computer” anywhere in the content.
  3. You can find experts based on the content they add; this is extremely helpful in large, distributed groups where people don’t know each other. By applying quality tags, and then correlating them with the person who added the content and the groups reaction to the content, we are able to associate tags with people who have authority on them.
  4. We passively suggest related content giving people the best possible chance to find what they need.  We do this by suggesting related content based on the semantic relationship of the tags.
  5. We flat out get better tags.  People tend to be lazy with tagging and don’t do it unless they are really committed.  We take this barrier out by doing it automatically.

Good tagging is key for enhancing content and making knowledge smarter.

P.S. I got the idea to write about this after commenting on another blog.  I thought, hell, I thought all this through, I might as well put it on our blog too.