In my last post on UGC and the learning professional, I wrote about quality.  This time around, I want to tackle accuracy.  Sometimes when talking about UGC and peer-to-peer learning, I hear an objection about accuracy – the content won’t reflect company policy or sanctioned procedures, or worse, federal or state regulations.  As with the issue of accuracy, there is a kernel of truth in these concerns, but it’s also a bit overblown.

First there is the issue of self-efficacy.  People who post “knowledge content” online rarely do so without high levels of confidence in their own understanding and accuracy regarding the content.  They also generally require high levels of self-confidence and a willingness to bear some level of scrutiny or criticism regarding their post.  Further, they need some sort of understanding of the medium and the technology.

The potential of employees posting inaccurate content, in other words, is partially self-regulating.  The more sensitive, controversial or “legal” the topic, the less likely “regular Joe” employees will be to comment on it, post about it, or otherwise generate content.  These self-limiting influences, coupled with the lack of anonymity in an employee community (and the resulting accountability), limit contributions on sensitive topics like sexual harassment, OSHA compliance or the like.

Second, most of the risks associated with inaccurate content can be mitigated through moderation technology and services.  Moderation should come in at least three flavors:  pre-moderation, post-moderation, and technology.  Live moderation for “live” events is also sometimes critical, depending on the type of community and the business purpose.

Pre-moderation is basically a screen to filter out content before it hits the community.  The user submits and then based on flags or as a blanket rule, the content is sent to a moderator for approval.  This moderator can be an internal SME or traditional gatekeeper, or external bodies employed just for this purpose. On approval, it’s put back into the stream to make its way to wherever it was going: blog post, discussion, comment etc…  This model makes a lot of sense in highly regulated industries like banking and financial markets, healthcare, or pharma.  It might also make sense on selected topics within a general workplace community – certain HR topics, highly regulated topics, etc…

Post moderation can be a combination of self-policing by the community or self-policing by SME’s and gatekeepers.  I purposely call this “self-policing” in both cases because SME’s and other gatekeepers (instructors, legal, marketing, etc…) are not separate from the community; they are a vital part of the community.  It’s not an “us and them” thing; it’s a “we” thing.  So whether “regular” members flag content as a violation or whether it’s done by experts, it’s still self-policing.  For many non-legal / non-regulatory topics, this kind of self-policing is sufficient to ensure content quality.  Like pre-moderation, post moderation can also be triggered through technology like word flags, filters etc…

Live moderation is often used in support of events where there is some sort of “live” interaction – live blogging, live chat, live Q&A.  Generally, this kind of moderation is a combination of carrots and sticks:  carrots = surfacing quality content from the stream to ensure a dynamic and engaging conversation and sticks = suppressing inappropriate content to maintain brand integrity or keep the conversation on track.

In all of the above scenarios, moderation services and technology provide a safety net to ensure that the content of the community is accurate and appropriate.  More importantly, they limit legal exposure from inaccurate content.  The only question is “On any given subject, how long can inaccurate content remain in the community?”  If the answer is “never,” then pre-moderation is required.  If the answer is anything other than “never,” then post-moderation coupled with necessary SLA’s can be the answer.

One thing that should be obvious in all of this:  don’t go it alone with a pure technology play or with “freeware.”  Getting started with Blogger or MediaWiki is not a bad thing.  Basing your whole strategy on these approaches without considering the bigger picture of moderation and content accuracy is a recipe for disaster.  That’s where a company like ours (Mzinga) can provide huge advantages.  As you can see, we’ve already thought about a lot of this stuff.  You should also be sure not to make the mistake of assuming all vendors do moderation in the way I described above – most do not.  Many lack real moderation services and most lack the technology back-end required for either pre-moderation or SLA-based post-moderation.  So do your homework.

The final points to make about accuracy in user-generated content are scale and delusions of grandeur.  Why scale?  Here’s the thing:  there are 80,000 articles in Encyclopedia Britannica; there are over 2 million US articles in Wikipedia.  So which is more accurate?  While it might be debatable on articles 1-80,000, there is no debate on articles 80,001 to 2,000,000.  Wikipedia wins because there is no point of comparison.  In other words, user-generated content provides a level of scale that cannot be replicated in expert-vetted systems.  Training, like the traditional encyclopedia market is expert-vetted and even expert driven. So are traditional knowledge management systems.  Social media, by contrast, is user-driven and thus relatively scale-free.  This is why Intel was able to generate over 200,000 articles in just under 2 years through an internal wiki.

The final reality to consider is a stark one.  Time to face some hard truth – most of your “experts” aren’t nearly as expert as we’d like to believe.  For many years, I was involved in systems training – Oracle, SAP, PeopleSoft rollouts, stuff like that.  I would go in as a snot-nosed kid who had never seen some of these applications in order to build training for them.  Invariably, as I learned the system and made assumptions about how the functionality worked, I would end up teaching the so-called SME’s all kinds of new stuff, not just simple things like new shortcuts, but whole new ways of using the system.

The reality is that many SME’s become institutionalized – in other words, they are identified as SME’s because of their expertise – usually ground-level expertise – and then they become this mythical being called a “SME.” Unfortunately, as soon as they achieve this exalted, God-like status, they begin their fall from grace.  The problem is that once they become a SME as a part of their job, they do less and less of their “real” job and thus have less and less connection to the ground-level expertise that made them a SME in the first place.  Throw in lots of business change, new technologies, new process, and some months or years, and the SME may not be, or at least not as much.

Assuming you agree with this, the solution is to ask everyone to be a SME.  We’re all an expert in something.  Joe, in accounting, may not be a SME in the broad sense of the term, but he might be the best person internally at reconciling accounts that are 90 days past due.  Mary in sales may not be the best sales person, but she might be great at cold calling or starting up dialog and conversation.

The big idea in all of this is to think of the whole organization as a collection of SME’s.  An even bigger idea is to help them to think of themselves as SME’s and for the organization to mine its own expertise.  Your role in all of this?  Change your thinking.  Stop looking to SME’s and start training and empowering your whole organization to be SME’s.  Stop “teaching a man to fish” and “start teaching a man to teach.”  Stop looking to the anointed few who have been dubbed SME or expert by the king and start looking to the wisdom of the collective many.

Next week?  Approvals.