Info Image

Big Data Disrupts Telecom Analytics, Fuels Contextual and Real-Time Customer Engagement - Flytxt 2015 Predictions

Big Data Disrupts Telecom Analytics, Fuels Contextual and Real-Time Customer Engagement - Flytxt 2015 Predictions Image Credit: Flytxt

ONE-CLICK TWEET

By Abhay Doshi,

VP, Product & Marketing,

Flytxt

Big Data has become the new lifeline for Telcos across the globe to reverse their declining revenue trends and preserve the customer base. It has been proved that significant value can be extracted from data provided right analytics solutions and business models are in place for internal and external data monetization. The input data is huge, diverse and unpredictable. Business users’ demand for processing these streams and feeding intelligence much faster to myriads of downstream systems and business workflows to extract optimal value within the window of relevance. Big data is disrupting the way in which analytics solution are designed and delivered in the Telecom world; let us have a look at some of these disruptions already visible in the market.

#1: DATA STORAGE, RETRIEVAL AND PROCESSING IS NO LONGER HARDWARE INTENSIVE

Crunching huge volumes of data is not something new to Telecom enterprises. However with the maturing of big data technologies and frameworks like Hadoop and No SQL, architects can now put in place a much faster and a massively distributed, scalable infrastructure to store and retrieve data as and when needed for processing. They are no longer dependent on highly inflexible and costly hardware infrastructures to achieve desired scalability in the big data era. Even the computing power can be scaled up and down on a need basis utilizing cloud technologies and the emerging intelligent layer of software defined computing.

#2: COMPUTE AND EXECUTION IS NO LONGER DONE IN SILO

Hadoop is emerging as a general purpose operating system for architects to manage massive data sets from myriad of sources of varying type and form. Gone are the days where enterprise systems had standalone components in the architecture to process these varied data streams. Now with Data Lake, you can store the data in native format itself. Multiple data transformation and processing workflows can be overlaid over the same underlying storage infrastructure to get the data extracted, transformed and loaded on time for downstream applications. The coming year will see most of the progressive Telcos attempting this irrespective of myths still existing around the feasibility of having Data Lakes and hybrid processing.

Abhay Doshi,
VP, Product & Marketing,
Flytxt

#3: DATA TO INSIGHT CYCLE IS NO LONGER TIME CONSUMING

Extreme real-time computing is arguably the hottest technology in demand for application providers in the big data analytics space. We are seeing significant partnerships and integrations to advance big data platform capability in this direction. Data from the underlying MDM layer and events from streaming sources like telecom networks and sensors needs to be transformed to insights in real-time. The near real-time is slowly moving into real-time and we will see more use cases of streaming analytics like contextual touch point recommendations and real-time event triggered customer engagement.

#4: NEW APPLICATIONS NO LONGER MEAN MAJOR CHANGE OF INFRASTRUCTURE

Telecom enterprises have multiple problems to attend to, which demands different analytics applications and decisioning workflows. Building business applications in big data era has become quite fast as the underlying data infrastructure and analytics layer are getting reusable either fully or partially. This means applications can be deployed without much changes to the base infrastructure; it can integrate all and any kind of data and run different kinds of analytics algorithms ranging from descriptive to exploratory, predictive and prescriptive. Business logic can reside on top which can leverage different analytics models depending on the kind of intelligence needed, and these analytical models can in turn pull in required data from underlying MDM layer with ease. It is a top-down approach when it comes to designing solutions as against the conventional way of building everything from scratch, thus reducing the time to add specific business applications.

#5: LEARNING IS NOW NOT LIMITED TO WHAT HUMAN EYES CAN SEE

Earlier there was a limitation on how much data can be analysed as it involved human experts. Advancements in cognitive computing, spatial visualization, and so on, is making analytics deeper, where machines can automatically derive learning and patterns from huge data sets analysing them in multiple dimensions. Now not only the whole data can be analysed but in all its depth too, bringing much more diverse and deeper insights, the models get automatically tuned for accuracy with enforcement learning over a period of time. This would throw up insights which could not have been possible with just human eyes; these deeper insights bring new opportunities for value creation too.

#6: ANALYTICS AND INSIGHT DISCOVERY NO LONGER REQUIRES ADVANCED STATISTICAL AND TECHNICAL KNOW HOW

Analytics applications and platforms call for role driven layered analytics and an integrated approach where data scientists create algorithms, domain specialists package these algorithms into reusable models and end users (like marketers) consume these models for enabling faster decision making and value creation. Self-service, interactive platforms for data discovery, visualization and analytics would find more takers in the market which will allow business users to play around with data and fine tune the packaged analytical models without technical and statistical expertise to meet their niche needs.

#7: BIG DATA MONETIZATION NO LONGER MEANS CONSUMER PRIVACY IS COMPROMISED

Analytics driven insights have got significant value when it is shared with third parties to drive new business opportunities like mobile advertising, location-based services, etc. The one factor that has led to the failure of many of these projects and initiatives of Telcos is the concern on the privacy of customer data. This seems to be slowly moving out with data security frameworks and Personally Identifiable Information (PII) to non-PII technology getting matured. These technologies ensure that any personal information is removed and incoming data streams of a customer are transformed to aggregated views and abstracted insights which protects the data and consumer privacy completely. Analytics platforms need to have this capability built in to support external data monetization.

Big data is disrupting the telecom industry in many ways; more importantly it is bringing business users to the spotlight. They would start playing a more significant role in determining not only the functional roadmap of analytics applications, but also the architecture roadmap in the coming years.

About The Author:
Abhay is responsible for Flytxt's Product Strategy, Product Management & Marketing. Prior to joining Flytxt, Abhay was responsible for Product Management & Marketing of 3G and 4G products at Huawei Technologies, Shenzhen, China. Earlier in his career he was involved in Planning and Engineering of networks & services for Reliance Communications, India and Product and Services Consulting for WANA, Morocco. Abhay holds a Master’s in Computer Engineering from the Florida International University (FIU).

ONE-CLICK TWEET

PREVIOUS POST

2015 to See Operators Shifting to Dynamic Charging Schemes Enabled By Virtualized Architectures - Qosmos

NEXT POST

More Pay-TV Operators To Join the OTT Frenzy as Demand Soars for Multi-Screen Video Content