Social Networking Analysis

Social Network Analysis is the study of social  networks and their application to solve real life problems.  In my graduate student days long  time back, I had  spent some time studying network marketing organizations like Amway.  One of the ideas emerging in those days was that exploiting ones network for business has some  positive externalities, due to which people get additional non monetary benefits, and were therefore willing to work for a lot less. After all these years, got back into this area largely due to the drive and interest of my colleague Devesh.

Modern day applications of social network analytics are more sophisticated.  Apart from the  HR problem I shared  in my previous post, here are a few other interesting applications:

TELECOMS --  Social Network Analytics can be used to manage customer churn. Basically an individuals churn likelihood goes up if someone from his/her network attrites.  Consequently it is more important to retain individuals who can influence more people through their network ( more on the quantification in a later post). This is one application where social network analysis  does not depend on social media data but rather on usage data.

PEER RECOMMENDATION ( Netflix type problems) -- For  situations where a business is trying to recommend a product/service to a customer, leveraging the preferences of  the individuals social network could come in handy.  This is very relevant for cross sell and upsell type marketing campaigns.

COUNTER TERRORISM  ---  Law enforcements authorities are using Social Media data to develop social networks and track terrorism suspects using these networks.

Here are some commonly used data sources:

• Social Media
• Customer Usage data (telecoms)
• Purchase Pattern
• Survey

For commonly used tools for doing social network analysis check this link on wikipedia

Here is a good reference on the topic Social Network Analysis (Quantitative Applications in the Social Sciences)

HR Analytics - an application

Following  up on my earlier discussion on HR Analytics, wanted to share an experiment I had participated in earlier this year.  A survey was sent out to  all our employees asking  questions related to people who  they sought out for technical problems, work related  problems and HR problems.  We then used to  survey to generate a mapping of the  employee network and created a list of who were the most valuable employees in term technical expertise and for resolving work related problems. We used social network analysis  to do the exercise. Want to also call out Devesh for doing all the analysis on the project. Watch out this space for a write up on social network analysis.

I see a lot of potential for this kind of application. Organizations can figure out the  network for each employee and  also get a measure of their value to the organization. This can have direct implication for employee retention. First you target your more valuable employees for retention, but also you can update the likelihood to churn scores for employees based on  events in their networks. One  shortcoming of the approach we  adopted was that it is survey based, so there is a possibility for smart  people to manipulate/influence the results.

Appreciate your comments..

-- datamining_guy

HR Analytics -- Is it going to be the next big thing?

As businesses get accustomed to using analytics to manage their customers, retain them, sell to them and acquire them,  they have started pondering if it could be used to manage their own people better.

Recently have heard several such comments/queries.

  One client recently wondered whether he could use empolyee performance ratings and resume characteristics to develop  a screening model to make life a lot easier for them.   ---- I would say go for it. Cool Idea!

Another client while collaborating on customer churn model wondered if same could be applied for employee churn. I think this can be done as well and will be immensely valuable.

I  think an increasingly  larger number of organizations will start using analytics to mange the employee life cycle better, and that is so exciting.


I would appreciate your thoughts.

datamining_guy

Real Time Analytics --- Foursquare and CNN

Progress  in analytics comes from two primary sources, (a) new technology  or (b) new data. The  emergence of real time analytics is probably a combination of both.  I have been personally very excited about the potential of  social media sites like  Foursquare for real time marketing.  For those not familiar with the application, it allows you to check in to your location. Therefore based on your location, you could  potentially receive specials on a  restaurant you are walking by or allows a small store owner to  give you some customized deals based on   your profile /comments.  To me that  is one of the biggest  potentials of this and similar innovations, the "democratization" of analytics. Allowing the smaller businesses to leverage the power of analytics to make better decisions.
In this context was very excited to see the announcement that CNN and Foursquare are teaming up to promote healthy eating by creating a "Healthy Eater" badge that can be used at locations like Farmer's market.

Data Mining Competitions

Folks,

One of the fun things for an analytics junkie is  participating in data mining competitions. Every summer there is the KDD CUP  followed by the  DMA  Analytics Challenge.  For folks wanting to do  year around competitions, check out http://www.kaggle.com/ , a web portal for data mining competitions. The latest buzz in the analytics competition community is the  $25k Hearst Challenge, which  launches on October 14, 2010.

Cheers,

datamining_guy

Hello

Hi  Folks,
Have been doing Analytics for over a decade and love what I do.  I immensely enjoy sharing my thoughts with  other folks and learning from their perspectives.  I also enjoy writing, so thought would go off  the beaten path, and do what everything else is doing ... start a blog :).

Promise to update this regularly and keep you posted on my thoughts, events,  interesting articles,  tools techniques etc.

Regards,

datamining_guy