Friday, February 23, 2007

Let's call it extreme predictive analytics

At our last Web Analytics Wednesday in Quebec city, we had the chance to meet with the cool guys working on Swammer, the result of years of scientific research. The step from what we know is so huge, at least for me, that it's a bit hard to really convey the power and the impact of what this could mean. So let's proceed with a couple of building blocks comparisons.


  1. web analytics, as we know, is the measurement of the behavior of visitors to a website
  2. predictive analytics process current and historical data in order to make "predictions" about future events
  3. KPI are metrics used to quantify objectives to reflect strategic performance of an organization. In the Swammer context, the KPIs are: Visibility, Actuality (who is more "vocal"), Celebrity and Popularity.
  4. reputation management is the process of tracking an entity's actions and other's opinions about them.

Advanced concepts

Still fine, isn't it? Now some new concepts:
  1. Key Genetic Indicators (KGI) maps de DNA of an organization, the social values and representations they are associated with (ethic, empathy, flexibility, efficiency, etc.) and their counter values. This is based on axiological and socio-psychological concepts.
  2. Key Tracers Indicators (KTI): are synthesis indicators derived from KPI and KGI to provide an interpretation of a social value (an image) across time. So KTI gives a representation of an object (a company, a product, someone or anything else) in 4 areas:
    1. the inner self image representation
    2. the projected image
    3. the perceived image
    4. and the imposed image
Each of these images are quantified to get indexes such as disparage (who is saying bad things about a company or a product?), conscientization (who says what in positive terms?), redundancy (how many occurrences of a positive or negative value?), etc...


  • The data set for predictive analytics is the whole visible Web, but could as well apply to other data sets
  • There's a way to establish the trends, think about a Google Trends for social values
  • You can establish the relationships between the social values and counter values associated with each element being analyzed. Something like AmazNode at a much larger scale
  • It goes beyond business intelligence, although similar in some ways to what companies such as Resilient Corporation or Cymfony are doing, Swammer takes a different angle and goes far beyond them


  • Marketing: You can get, spot on, the values and counter values of your competitors. This become a very strong tool for SWOT analysis. In this example, banks, pharmaceutical, oil companies, car manufacturers or any industry can be analyzed.
  • Political: It as been found that often, official election results are surprisingly close to survey results conducted close to the election date. Imagine if you could get those results weeks before what traditional methods allow today. Swammer is being tested for the current elections in France and in Quebec, and could certainly be used to see how Obama and Hilary Clinton are doing...
  • National: Values and counter values found in our society leave "traces" and "artifacts" on the Internet. If you can monitor those, you can spot specific activities... such as terrorism. Some groups have a constant "buzz" on the Net, but sometime, a sudden "blip" on the radar from an obscure group could hide some emerging activities. Although they could not provide too much details (go wonder!), the guys from Swammer revealed they have been approached by secret services to see how their concept and technologies could be used.
  • Legal: demonstrate such things as when and how, in case of legal litigation, a company as been a victim of defamation or other socially identifiable counter values.
  • Financial: One of the next research field will be to use the concepts highlighted above to predict future activities on the stock market. I kindly volunteered myself to test this one :)

Full disclaimer

I met the guys from Swammer in person during our WaW event, where they presented their technology. This post is partially based on material provided by Swammer. I do not receive any financial or other professional or personal benefits.