Garbage In, Garbage Out

  • Excellent post Christian which touches on one of the biggest challenges facing Yammer & Other ESN’s!
    I think it’s important to remember that very few organisations have trained employees consistently in how to use email, and it’s the exact same issue facing ESN vendors. It’s the opportunity that most organisations have been waiting for, and wipe the slate clean.
    With so many people raising the same “how to avoid moving the mess” question it’s time for all ESN vendors to step up. The fact is that people are generally lazy, and the belief that social tools solve information overload is a fallacy.
    One way I think Yammer could improve could be through implementation of auto/predictive tagging features. When setting up new groups you simply define a few basic tags which are applied to all posts/content within that group. Then extend this further where additional tags are suggested based on existing posts etc.? It’s not dissimilar to the Google “Did you mean” or Amazon’s “Other Customers Bought”.
    Extend this yet again by indexing/crawling the contents of documents, or a photo’s metadata to help surface a plethora of tags, both existing and new. Finesse is needed of course to avoid over tagging which would effectively render conversations and content useless.
    Posts & content without tags is yet another conversion metric which community/group managers need to take action on to prove the value of their ESN.
    On a positive not though, this is also a great way to score groups/communities based on their tagging prowess. “Tag Sprints” would be a great way for teams to end their week, where they spend an hour or so tagging the conversations & content facing an identity crisis.
    If ESN really is superior (which I think it is), and the vendors can solve the classification challenges in a more compelling fashion, then I think people would be more willing to adopt. We cannot automate everything though, and people still need to be willing to learn how to use these new tools.

  • Thanks for the feedback, Greg. I think there is tremendous value in ad hoc collaboration without the burdens of overhead and structure. I helped create a SharePoint-like platform for E2open 12 years ago to do this somewhat, and in the early days of SharePoint it was also viewed as a way to quick collaborate in a somewhat unstructured way…which is part of the reason why it grew so quickly. And then as it matured, it needed to become a much more structured platform. I think Yammer needs to remain largely unstructured — but have the ability to plug into the taxonomy, the content types, etc so that it can work with the structured, when necessary. The hard part is to keep it simple enough that you don’t over-complicate the product, and as a result, drive away end users.

  • Pretty much agree with Greg’s sentiments. While enterprise IT has done a decent fast follower act with respect to consumer IT from a social interface standpoint (variations of the Facebook motif) it has lagged under the covers in intelligently applying the social data that is generated.
    Such intelligence still requires sufficient behavioral information from which to make its inferences, so users need to do their part, to your point Christian. But the system ought to be doing a lot more heavy lifting. For example, let’s take tags. Whether Yammer or SharePoint, in their native form they are structured to primarily enable humans to do simple look-ups based on individual tags. But the machine can do much more sophisticated things with these humble tags. It can calculate tag frequency and apply the insight that infrequently used tags are more informative of affinities between content items than frequently used tags, everything else being equal. It can also keep track of individual usage patterns with respect to tagged content, and infer affinities between users and tags. (And cross-contextualization of tags between Yammer and SharePoint can be *automatically* learned from usage patterns between the two platforms.)
    Add to this the insights that can be derived from comments, likes, ratings, following patterns, etc., and it is easy to see that what Amazon and Google do in the consumer world (and more!) can and should be delivered in the enterprise right now. And I happen to know one company that is doing so . . . 🙂

  • Steven, you’ve just touched on what I think is the major disconnect between social tools and enterprise content management: how metadata is applied and then utilized. Most users of social platforms ignore the long-term value of conscientiously tagging content and conversations because of the future value of their actions, and instead focus on the immediate need to chat and share. Maybe that’s where machine-based tagging needs to come into play — automate what we can’t seem to do manually, thus allowing us to begin identifying all of those patterns, and get more value out of these platforms.