Today’s blog entry is again brought to you by me, John Hand. As the Senior Marketing Manager for Knowledge Network I regularly give formal presentations either in person or via another one of our great products, LiveMeeting. And one question that keeps coming up is, “How does KN compare to expert location systems?”  There are some subtle and some not so subtle differences. First the not so subtle differences. Many, if not all, expert systems require manual input. We have one here at Microsoft which I am hoping will soon be made obsolete or at the very least fully supplemented by KN, but that’s another story. Stay focused Johnny, stay focused.  Okay, so I mentioned the manual inputting part right? Guess what? Most people (including you and me) are lazy. That’s right folks. Given the opportunity to manually describe what we do and who we know, or to avoid that work altogether, most people --believe it or not-- would avoid the work. And so when an expert system is put into place and an edict is send down from above that “Thou shalt profile Thyself” many people enter into the task somewhat begrudgingly.  And so there you sit in front of a system with a cryptic user interface with the challenge of not only describing what you know but perhaps even who you know.  I’ve been here at Microsoft for 18 years and have had a variety of experiences. If I were faced with the difficult task of describing what skills I have I would be hard pressed to remember all of the projects and people that I’ve worked with over the years.  My degree is in code writing but I’ve never written a single line of code professionally at Microsoft.  So in my profile should I write computer programming?  Probably not.  Marketing, sure. Management, yes. Other stuff, events, TV experience... I’ve done so many different things it’s hard to remember them all.  That brings up one of the first short comings of these manual systems, recall. As beautifully designed as the human brain is, it is not very good at complete recall.

 

Then the social network part comes into play. Who have I worked with? We’ve all had that awkward moment when we see someone we’ve worked with in the past and they were nice enough to remember your name but, guess what? Gulp, you’ve forgotten theirs!  Relax, this is normal. If we ever meet face-to-face I expect you to remember my name. It’s John by the way.  So here we are faced with the second shortcoming of manual systems. What is the extent of our social network and how can we possibly remember or keep track of everyone with whom we’ve worked?  Again not to belabor the point, but in 18 years with Microsoft I’ve had the chance to work with many great people on various projects.  So even on a good day where I was feeling full of myself and you asked me to list my first-level contacts, what we call "colleagues" in KN, I’d have said that maybe I have 100 contacts.  Well I just switched over to my profile manager in KN and as of right now in the real honest to goodness Beta 1 of KN I have 237 colleagues and 228 external contacts (one of them might be you).  And if you made me sit down and name each and every one of them I’d be sitting for a very long time.  I would not be able to enter all those names into a manual system given that I don't have perfect recall. Yet, KN can figure it out with my only involvement being during the review process where I get a chance to review the automatically created profile.  Perhaps 2-5 minutes investment of my time, tops. 

 

Here’s the other problem with manual systems, assuming that you get all members to profile themselves first, now you have a nice neat database right?  Search against it and you get accurate information, right?  Well, maybe. But, my assessment of my programming skills may be entirely different than other’s views of my skills.  I actually own a book about C# programming but can I program in C#,? Sadly no, not yet anyway. But once I read that book, watch out.  So now we have to deal with disparate ways in which people assess themselves. Again we get back to human fallibility, this time regarding how we try to apply our assessments equitably and fairly across and entire organization or division.  And so the data is suspect from the beginning.  Plus we never move to other projects, right?  We stay in the same jobs forever, right?  We never switch roles or change divisions or companies, right?  So there is the second challenge. Keeping these profiles updated is an ongoing struggle.  And guess what?  People generally don’t do it.  So after the first nice and neat database of suspect information gets published it very quickly becomes out dated and the people using the system know that.  So we stop relying on the system. Since we know that no one else is taking the time to update their profiles we stop bothering with updating our own profile.  And then the system collapses under the weight of it’s shortcomings, GIGO, Garbage In Garbage Out. 

 

Lastly, assuming in a Utopian world that everything goes right with an expert location system and lo and behold you’ve been able to identify the expert in your organization on say, desktop deployment. Great!  All’s well with the universe. Guess what happens next?  That lucky individual who is deemed “The Expert” has a target on his back is quickly in demand and one of two (and often both) things happen. Either that person is quickly overwhelmed by the amount of requests and the queue to talk with that person gets longer and longer.  There are no provisions for finding people who have enough skills to help you so everyone else is in effect categorized as “The Great Unwashed” when in actuality there may be some very qualified people who could also be of great help. But not in an Expert Location System, it is just that one person so burn out can quickly occur. So pity the poor expert, and hope that you never make it to the top of the pyramid.

 

 

                                                       

So now that I’ve gone on and on (and on) about issues with “Expert Systems”, let’s talk about how KN addresses those problems.  First the profiling problem. Since KN automatically generates a profile for each member from their e-mail it automatically extracts your keywords including noun phrases and your key contacts (both internal and external).  Then once the profile is generated you have the chance to review it. You can add or delete any elements in your profile.  So the tedium of profile generation is eliminated and replaced with profile review, which, honestly, I’ve never taken more than 5 minutes to review my KN profile.  So the comparison is hours or days of error prone tedious work compared with 5 to 15 minutes of profile review with current up-to-date information of who you know and what you know.  So that solves the first problem of getting critical mass into your people finding system, because that’s what an Expert or “Expertise Location” system is all about, finding the right people vs. finding documents.  In fact I use the tagline, “It’s all about Connections” when describing KN.

 

So the second challenge is keeping this information fresh. Again, KN to the rescue.  Once you publish a profile in KN, the client software is continually collecting new information about what you know and who you know.  Then periodically (the default is every 14 days) you receive an e-mail that shows the net new contacts and keywords you’ve used in the last two weeks and sends you an e-mail with that information broken out by keywords, contacts and external contacts.  If you approve of everything you see you simply click on the “Publish All” button right inside of the e-mail and the server is automatically updated.  If for some reason you need to edit this new information you simply double-click on any of the items you wish to edit and you are immediately taken to the profile manager where you can make any appropriate changes.  So the freshness of data is automatically taken care of. So, problem two is solved.

 

Lastly, how do we address the issue of Expert burn-out?  Well, our system is not really designed to find “The Expert” but rather it is an organically created system for finding knowledgeable people, hence the name Knowledge Network (clever huh?  I came up with that myself!).  We help you locate people who can help you make a decision or solve a problem without having to designate one particular individual.  And in this way no one has to ever wear a target on their back.  So bye-bye burn-out and so-long long wait times.

 

 

Okay, I’ve gone on long enough. I hope if you’ve stayed with me this far, you now appreciate the subtle and not so subtle differences that KN brings to the world of expertise location.  I welcome your thoughts on the topic so join in and someday you may be one of my first level contacts.  Remember that in today’s business “It’s all about connections!”