The first step to micro segmentation
Micro segmentation simply means identifying groups of entities that have homogeneous behaviour. The process depends on both quantitative and qualitative attributes and to be able to compete with other service providers, it becomes necessary that telecom operators know enough about their customers.
Telcos have been working on segmentation for many years and there have been good many tools that have aided them in this endeavor. These often look at static parameters and then segment customers on the basis of these parameters, for instance age, spend, tenure, gender, services used etc. This was quite useful and sufficient to meet the needs of campaigning in the days gone by. However, with the increase of business velocity and the need for relevance becoming more and more critical, it is now observed that large segments do not reflect the need of the business. This has led to the development of micro-segmentation and obviously this calls for deep understanding of the data and how to identify narrow segments of entities that have similar behavior or preferences. For instance, it is not just enough to know customers that spend a particular amount each month, it is also necessary to know on what and when they spend and these would then become micro-segments.
How does Pelatro fit into this?
As seen above, the ability to micro-segment calls for a deep understanding of the data and it also calls for agility in handling this data. Pelatro has developed a solution, mViva, which has the ability to generate such micro-segments through the use of heuristics and machine learning techniques. These micro-segments are dynamic and even have the ability to go to an N=1 level. This is achieved through the use of the latest approaches in Big Data Analytics and by bringing in new innovation into some older techniques like hypercubes.
mViva is built on Hadoop, HDFS framework and is capable of sifting through billions of transactions on a daily basis. These are then classified and stored for future use. Such data can be used for long periods to generate accurate trends and these can then be used for defining micro-segments.