The statistics of content growth are truly awe-inspiring, whether measured in the number of bytes stored or the number of objects managed, the shear volume of content continues to increase and it is this volume that has brought about the entire field of ECM. One might believe that knowledge is growing at a similar rate as content, after all content is information and all information is knowledge therefore all content is knowledge, right? There is certainly nothing wrong with the logic of that sentence, however information and knowledge are really subjective concepts whereas content is an objective concept and so the logic is misapplied ... linking the two is convenient for both ECM vendors trying to sell the idea of a large repository and CFOs/CIOs trying to reduce costs and increase collaboration (bringing a windfall in new knowledge-based ideas) but wanting something isn't the same as having something. In fact, the connection between content and information or knowledge is often tenuous at best, and we all know that intuitively.

It seems pretty clear that there is a linkage between knowledge and content, but far from being the same thing, some content can be the antithesis of knowledge ... some of it contains misinformation, stale information, even disinformation at times. Despite providing some method for providing feedback from the consumers or users of the content, ECM systems are not capable of differentiating between good versus bad content or good versus bad feedback. Humans with the appropriate expertise (and knowledge) have little difficulty in determining good from bad and most people look to the feedback from such experts to help them distinguish knowledge from plain old content. This is not unique to information, of course, we look to expert opinions in everything from what movies we watch, what stocks we purchase (or sell), even what team will win. And the problem inherent with expertise sharing is in determining which experts to listen to. 

This is not an easy problem to solve and, in fact, I believe that most ECM systems as they are deployed get it wrong. The essential ingredient in almost all feedback algorithms is volume ... the more people indicate a particular person to be an expert the more we believe them to be an expert. As most of us are capable of recognizing a real expert, they are often flagged as such in these systems, however that same system is easily gamed and many wannabe experts are falsely promoted. The basic statistics cannot and should not be blindly trusted, rather the feedback itself should be scrutinized for biases, but this is just not done by the systems or the people using them.

The politics of most organizations suggest that internal feedback is tainted and untrustworthy. Who dares to suggest that content created from the boss is less than excellent? In fact, can we even trust that the content being attributed to a boss is in fact the bosses'? It is certainly not the case for external, public content where every word is usually written and carefully vetted by others until uttered by the boss. Even the President has speech writers. False attribution probably doesn't exist internally within an organization to the same extent as externally, but it is highly naive to believe that it doesn't frequently occur.

Feedback that works (in that it promotes knowledge sharing) must be trustworthy, to be trustworthy the person providing the feedback must feel free to be honest without any retribution. But let's be honest ourselves, shall we? How many organizations can truly claim to have such a place? Most people will keep negative comments to themselves, and many will offer undeserved praise. You can see this in action in almost every solution that provides feedback, once an information object is published, the immediate underlings of the publisher will "Like" or "Recommend" the content almost immediately. This is often done at the behest of the publisher in order to skew the results of what the feedback is intended to do in the first place.

And the person offering any negative comment must themselves be able to provide better than average work or else the feedback to them for that negativity will be fairly strong (it will be anyway, the person making such a comment must be very sure of their position within a company, or too naive to understand the consequences). In general, though constructive negativity is in fact far more valuable to an organization that 10 times as much praise (deserved or not), it is very difficult to obtain in a public forum.

Anonymous Feedback Only Partial Solution: Ranking versus Rating

This same problem of false positives exists within places that offer anonymous feedback, in fact the gaming of the system is wide-open to abuse with anonymous feedback ... to see this in action one only needs to spend a few minutes on the Internet looking at customer reviews of nearly any product or service. Many of the reviews are clearly dishonest attempts of gaming the search engine ranking heuristics, their contents are less important than their existence.There is a big game going on between Google and many web site owners where Google is constantly trying to rework it's system to catch a notion of rating as well as ranking whereas many web site owners are trying to beat the system with statistics rather than content.

But anonymous feedback does allow for a less encumbered and perhaps more honest review.

If the person providing the feedback knows and believes that their comments cannot be traced back to them, and that their feedback might bring value to something that is important to them, then they are more likely to provide constructive criticism. Unencumbered with concerns of retribution along with their desire to improve the thing they are providing feedback about, the feedback can be highly valuable. But of course that anonymity is also the Achilles Heal of anonymous feedback.

The human capacity to inflict damage upon others is almost infinite. Where one person / group is elevated in recognized expertise others are demoted similarly ... it is a zero sum game as the name suggests; it is a ranking not a rating. If I become the number 1 expert on an existing subject, then the previous number 1 expert is demoted to number 2 and so on down the line. In fact, there is no way for me to elevate to any particular expert-level without taking something away from somebody else, such is the nature of ranking. It is not hard to see that it is often possible to be more of an expert by simply downgrading another's expertise rather than doing anything proactive. Particularly if the negativity cannot be traced back but as political attack ads can attest to, this equation holds true in other situations as well.

Thus anonymous feedback encourages false negatives whereas attributed feedback tends to limit or even eliminate (false) negative feedback.

Is there really an answer out there? Yes, definitely, but not for everybody.

With both attributed and anonymous feedback providing unsatisfactory results, what can be done? I think the average person intuitively knows how these things work, it is based upon trust. We listen to movie reviews because we believe previous reviews were accurate ... we TRUST the reviewer as an expert. It is easy to build up this trust, one only needs to go to a few of the movies being rated to see if you agree with the rating. Note that movie reviews are usually RATED and not RANKED and it is the rating that we trust (or not).

Many social networking advocates believe it is possible to capture this notion of trust and spread it out further. If you know me and like the movies that I like, it is probable that you will agree with my view of a movie reviewer and therefore trust that reviewer without first talking to me. The hope/desire for social networking, as it is being adopted by ECM vendors, is that that trust model can be captured and used to identify experts and by extension the knowledge and information that they find useful, interesting, accurate...or not as the case may be. On it's face it seems intuitively obvious that this should work but in reality it fails miserably at it, why?

It is possible that it is the solutions themselves that fail and I think they do play a role, but I believe it to be more fundamental than that.

What does "like" really mean?

Recently I found myself on Facebook pressing "Like" to a death notice, though I very clearly did not mean to suggest that I liked the fact that the person was dead. I didn't know the person but respected their life's work, and thought the tribute was well written and captured the person perfectly, so I "liked" the content. In fact the person themselves were related to my field of expertise and so somewhat related to this trust model we so want to capture ... I very rarely comment on the death of anybody, to worthy a "like" from me really said something about this person and that (some of) their work is useful (knowledge).

But then I also enjoy a good pun, joke, riddle, and / or giggle and "like" anything that make me chuckle, regardless of whether it is related to my expertise or not. I know with certainty that my humour is not shared by the majority so that "like" as it refers to humour simply shouldn't hold the same weight with most people as my expertise in computing might. 

So, exactly how is an automated system supposed to understand that my "like" was in fact due to the humour I saw in a situation that I otherwise find goes against my expertise? Does the fact that I am (hypothetically) rude enough to laugh at other people's misfortunes make me more or less an expert? Does my concern for how people might perceive my "likes" dissuade me from letting people know I laugh at other's misfortunes? And if it does then is the system really capturing the expert knowledge it claims?

More options are needed but won't be used

The fact is that "Like" or "Agree" or "Recommend", the opposite of each and any other single moniker are inherently too simplistic to capture the same sort of web of trust spoken about with the movie reviewer. One really needs a myriad of different buttons to capture one's feelings, one can legitimately like something and disagree with it at the same time, that isn't necessarily orthogonal. One could easily recommend something that they believe is a good example of what not to do, one could entirely agree with a comment but dislike it just the same. Comments go a long way to disambiguating the choice but that puts the burden back on the reader, we do not yet have anything that can pull the context out of a comment.

So why not a series of icons? Like, Dislike, Agree, Disagree, Recommend, Funny, Accurate, True, Sad, Happy..... well I guess the answer is obvious in the question, it is simply too complicated, a list such as this would likely stop  many people from providing any feedback whatsoever. The problem isn't so much that the list itself is too complicated, it is that the GUI would be too cumbersome, too confusing, too intimidating.

So what is to be done?

I think that small business can afford to experiment and, in fact, the smallness allows for an intimacy that makes social networking effective in exactly the opposite way that a large business tends to make people feel like a cog and social networking little more than the play-toy of a few. In small  places the feedback is inherently more honest and intimate, the mutual shared goals are real not forced upon you. Negative feedback aka constructive criticism is more likely and the false negative reviews much less likely to be expressed. This comes not because of any particular technology but because of a sense of "family" inherent in a well-run small business that simply cannot be found within larger groups.

Big business cannot build such an environment using ECM and/or social networking tools ... particularly public businesses that focus on the bottom line. There are too many goals, many of which conflict. Like 'basic research' the hoped-for benefits of knowledge sharing are all but impossible to asscertain. A provider of feedback has much to lose and little to gain and thus most often the feedback is limited to positive messaging because that doesn't threaten anybody.

There really is hope!

The key is in building out a community rather than centralizing and pidgeon-holing. It has to be real, the members must want to belong because belonging is the thing ... not because belonging brings them something. Only that community brings the honest and useful feedback and no technology in the world can build that community. Don't look to the tools to create your social networking/ECM goals, build your community and then implement the appropriate tools. If centralizing your ECM strategy makes economic sense, then take advantage of that economics, but don't throw out the baby with the bathwater. Build you communities within your ECM infrastructure, find your champions, your knowledge leaders, and encourage their interaction by removing as many causes for concern of retribution as possible. Allow for anonymous tagging and commenting but within a closed community...let non-community members view the discussions (if they want) but only allow members to post anonymously. At least then you can know that the group of people providing such feedback have shared goals and (probably) no ax to grind.