Category: Analytics

In my continuing coverage of KXEN, the plucky company that has managed to revolutionize analytics automation and social network analysis- Here is an interview with KXEN’s Vice President Bruno Delahaye.

Ajay – What is the best feature you like in KXEN. – both   as a company and as a product.

Bruno- Well actually what I like the most about KXEN is the will to make a difference. This is true at different levels of course: each individual within the company is trying to make things happen. For employees at KXEN this is not just a job: they want to change the game! The product side is naturally cascading from this. We are not simply recoding existing algorithms like some of our competitors are doing, instead we are looking in every domain of predictive and descriptive analytics where we can deliver higher value to our customers. When customers, thanks to the automation we provide, come back to us stating that they manage to increase their modeling productivity by 10 or even 50 compared to their previous modeling process we really think that what we provide is changing the game. Also, the fact that we have well over 500 customers globally today is proving that our customers recognize this as well!

Ajay : What areas has KXEN been most suitable for ? Biggest success story so far.
Bruno- KXEN has been very successful for 2 types of customers. We have been very successful in companies with mature Data Mining practices, companies that have realized that they need to move from a fully hand crafted approach to a more industrialized one in order to answer business requirements. As an example, lots of large companies run 10s of marketing campaigns per month and actually use data mining for only 1 or 2 at best… once organizations have understood the power of Data Mining they certainly want to target each campaign. Only KXEN can provide the level of automation required for this. On the other side, new data mining users (either new companies or new departments in a company) are also very eager to use KXEN. The learning curve with KXEN is so quick that it enables them to use their existing team (the ones that are aware of the business issues) and make them run within few days successful churn management programs or rebuild their customer segmentation in a reliable manner.

If you were expecting figures here, some Vodafone entities are claiming that they reduce churn in some customer segments by more than 10% by implementing KXEN. Unicredit in Austria mentioned that due to KXEN they gained an additional 50m€ per season….as you can guess the success of our customers always brighten our days.

Ajay : What areas would you rather not recommend KXEN? What other software would you recommend in those cases ?

Bruno- Well, I would recommend to use KXEN in every area of course, nevertheless where we have been less successful so far is with companies where time pressure to deliver analysis is lower. Basically, research departments tend to use more softwares like SAS EM or SPSS Clementine that are more methods/algorithms oriented rather than results oriented.

Ajay : What is the biggest challenge you have faced while introducing KXEN to a wider audience.
Bruno- The bigger challenge we have is in building domain expertise, it is indeed very difficult to build knowledge of our teams at the same time in Customer Lifecycle Analytics, in HRM, SCM… that is where building a confident relationship with the customer is so important. We have to prove to our prospect very early in the discussions that with KXEN they will make significant steps forward! This is also where our partner are so important to us. KXEN works with international as well as local partners with specific expertise to help our customers make the best possible use of the KXEN Data mining software to insure a high and fast ROI.

Ajay -Do you think the text mining as well as the Data Fusion approach can work for online web analytics, search engines or ad targeting?

Bruno- The data fusion approach is certainly one that makes sense for online web analytics. Analyzing the sequence of events rather than just taking into account whether an event occurs is actually a very powerful way to predict customer behavior or in this case the next click or the next action that is going to be made. I am not in this case claiming that everything has to be real-time as this could be the cause of the creation of weak or even unreliable/non stable models. Instead what we recommend our customer to do is to split the learning part that can be made off-line from the deployment that needs to be done real-time.

Ajay- Describe the relationships of KXEN with other members of the business intelligence community in terms of alliances.

Bruno- KXEN is a very good complement to BI vendors. We are actually partnering with several Data warehouse Vendors. For Data warehouse, the equation is quite simple they allow customers to structure and store the data but to provide real ROI, solutions need to be plugged on top of them. Setting a Data warehouse if you do not use the stored data is just another cost, what KXEN does is enabling to take advantage of the data asset to build customer segments that you will use to define your marketing mix, or simply target your customer either for cross-selling, up-selling or retention/loyalty purposes. The same is valid for credit scoring, fraud detection….

Case Study- Assume I have 50000 leads daily on a Car buying website. How would KXEN help me in scoring the model (as compared to other online based scoring solutions). Is it technically possible for me to install KXEN on Windows/ Other instances in remote computing like Amazon EC2 and not a server sitting somewhere.

The key difference, I believe, is that with KXEN you will indeed be able to do this even if you are not a data mining expert, if you want to use the results of yesterday’s campaigns to rebuild a model and if you can only afford to spend 10 minutes on this task every day. At the end of the day what we allow our users is to answer their business questions within the time frame they have rather than trying to convince them that they do not really need to do so many analysis for their business to run successfully

Ajay- And that was Bruno, VP, EMEA, KXEN. His profile can be seen here

http://www.linkedin.com/in/brunodelahaye

Bruno Delahaye manages KXEN’s operations for Continental Europe, Middle East, Africa and South America at KXEN. He is responsible for identifying and managing key partnership opportunities and developing the overall strategy for new partnerships.

For more on KXEN please go to http://www.kxen.com, you may need to regsiter to download their properietary white papers on Structural Risk Management or Text Mining.

Conflict of Interest Disclaimer-I am a consultant to KXEN as a social media consultant. Chairman Roger Hadaad was one of the first Chairman of a major corporation to agree to give interview to this small blog.