Tag Archives: IBM

Prediction Markets here to stay?

Web2.0 is all about contributions from end-users aka. The wisdom of the Crowds. Prediction Markets are a very efficient way to consolidate that wisdom which can be found in various nooks and crannies. If this concept is new to you, read Wikipedia for a primer. If not, read on.

Prediction Markets always receive a huge surge around election time because studies have shown that with enough activity they are more accurate predictors than polls. This year InTrade.com went very mainstream getting mentions on numerous news reports and getting link outs from Yahoo and other mainstream online information providers.

There was a move by DARPA in 2003 to use a prediction market to predict terrorist activity. While ideally that was a good approach to help understand the threat at any given moment, someone forgot to consider that most Americans don’t currently understand the concept of a prediction market and so just heard, “We are going to be betting on terrorist attacks”. You can imagine that program was shut down quickly.

So, will Prediction Markets maintain their momentum post-election this time around? I certainly hope so. We have learned from the worldwide financial crisis that we need people who make decision to have more information at their disposal rather than less. BI, Balanced Scorecards, and, yes, prediction markets should be getting attention/funding from those who feel that their organization is being led too much from “the gut”.

Continue reading


Social Tagging

Social bookmarking (eg. Del.icio.us) is one example of social tagging. IBM’s internal Fringe application (also called Bluepages+1) is another example that goes beyond social bookmarking to tag people rather than web pages. In this author’s opinion social tagging is the most valuable feature of a corporate whitepages.

Social tagging is merely the ability for multiple people to describe a single object with free-form words. It is far superior to author/creator tagging because it allows those who are making use of the item to be the ones who are describing it. Thus, the descriptors are likely to match the way that the target audience perceives the item and its value.

A pretty academic discussion of social tagging is available from a British journal called Ariadne or the more colloquial version is available from Wikipedia under the topic “Folksonomy”.

I will leave the discussion of free-form tagging vs. structured hierarchical taxonomies to another post, but having created structured taxonomies for multiple businesses in the past, I know there are many difficulties/drawbacks to doing so. Tagging is much more organic and scalable though it does have some drawbacks.

One of the most difficult elements of creator only tagging is that the incentive to tag is purely altruistic (good of the firm) or recognition based (I want people to find this so they know I am the expert). Social tagging introduces an additional incentive. In tagging something/someone, I am creating a breadcrumb back to that item. At any time I can recall the items I have tagged with a particular word. Also, not only am I creating a better sense of the item I am tagging, but my tagging act is helping to describe me. A view of my tagging words helps to understand me in the same way that perusing the bookshelf in my home would.

Even the number of tags on an item tells us something about that item. An item with more tags may be more popular. We have to be careful with this until social tagging is mainstream because there will be a skew towards the naval gazing of web2.0 type topics.

As with anything the social tagging user interface is ever important. If it is cumbersome to enter tags, users will not do so. The Fringe example from above was ideal because one generally looked up another user for their phone number and then while you were chatting to them could multi-task to attach some tags to them about the topic being discussed.

One of the most important factors in social tagging is the long tail. While many people will not social tag, there will be enough people in the long tail who derive personal and altruistic value from the act of tagging that they will tag a large portion of the items in order that the masses can better search.