Big Data effect on CSPs: Millions of segments to sell!

AbhayBig Data is here and now for CSP’s. You have lot more data coming in regarding what your consumer is doing or spending, thanks to the ubiquitous device, the mobile phone. Earlier marketing was more about doing the right thing for masses; it shifted to doing things for segments. Segmentation however was still more of an approximation where you group people based on their generic usage behavior or demographic profile. The one major con with such a segmentation approach is most of the time you end up in over selling or short selling to individuals with in a segment. It leads to leakage, lower conversion and more importantly customer dissatisfaction. It also explains why traditional flat usage plans or loyalty programs failed to deliver economic value for operators in the long run as they were mostly designed for broader segments.

big-dataMulti-dimensional granular segmentation
Segmentation accuracy is directly proportional to the number of attributes we use for clustering or grouping. With Big Data you have the opportunity to bring in more variables than ever before.  Besides demographic profile and usage behavior, you have lot more insights like product affinity, spending power, purchase history, psychographic preference, behavioral preference, time of day’s preference and handset profile. This will improve segmentation and make it more granular. Only telecom operators can claim to have such a wide array of variables to segment the broader population. Granular segmentation helps to reach the bottom of the pyramid to find the right group of consumers to be targeted for selling. Now operators can use this for selling their own products or enable third party to promote their product and/or services.

Fluid Segmentation
The consumers have become dynamic in nature as their needs and preferences vary from one moment to the next one. For example a subscriber who is on roaming may look out for good dining options nearby or would like to send few messages to near and dear ones back home. Once he is back in his home network he would switch to his normal usage mode which could be downloading music or checking out emails. Doing segmentation on hard boundaries and once in a week or month may not help in identifying those ‘moments of opportunities’ with consumers to sell. With smart devices penetration and mobile broadband growth, touch points and contexts of use have gone up exponentially.

Big Data Analytics can help to sift through these millions of contexts to identify the up-sell/cross-sell recommendation opportunities. It basically brings in dynamic or moving variables like ‘Those who recharged in last hour’ or say ‘People who switched to roaming more than X times last month’ or even real-time contexts like ‘those who are traveling to location Y’ into segmentation. In this way you can let consumers to be in multiple segments, for example he can be in business user basket in one context and can belong to foodie segment in another context that helps in identifying what to sell or promote to them. Normally marketers don’t get aggressive with high value customers, but some of those may be having low data usage, which is again an opportunity to cross-sell. These opportunities will be missed if we define segments only with hard boundary conditions. Fluid segmentation is also useful when it comes to dealing with volatile subscribers whose usage and recharge behavior drastically varies with seasons, economic conditions and even locations.

1-1   Segmentation
Segments are giving way to 1-1 selling. This is made possible thanks to advancements in real-time event handling and auto recommendation engine. As it is now possible to pre-build persona of each subscriber reflecting his interests, preferences and contextual needs, it is  now possible for a machine to recommend rights products/offers to be promoted to each subscriber on a based on predicted  context. Auto recommendation engine works on predictive analysis principles, but subscriber intelligence powers the engine to make the right choices.  When two consumers recharge with same amount, different offers can be sent depending on whether the person is frequent SMS user or a music fan who loves to download ringtones. While predicting the subscriber persona, it is important to keep in mind that a subscriber has graded affinity towards same products or different products at any given point of time. Degree of affinity will be important to take an appropriate choice.

It is complex, but effective
As operators move from hard bounded single dimensional segmentation to multi-dimensional fluid segmentation and then to pinpointed targeting and 1-1 segmentation, the sales effectiveness is proven to increase. You reduce the possibility of over selling or short selling as you know consumers more in depth and you are reacting to his needs in real-time.  Does it sound complex, it may, but now with man-machine collaborated decision making, powered by Big Data Analytics technology platform, operators are already moving to adopt these dynamic segmentation approaches. They have now millions of segments now to sell effectively!


By Abhay Doshi, vice president, Product and Marketing, Flytxt

Also Read

Leave a Reply

%d bloggers like this: