Bye Bye, Big Data – Hello Intelligent Fast Data

Posted: April 2nd, 2015 | Author: Mikko Jarva | Filed under: Industry Insights | Tags: , , , | 1 Comment »

For some digital and communications services provider executives, the Big Data trend has been a big disappointment. Operators were entranced by the idea that rich data analysis can reveal targeted insights that drive more revenue, but not every telco has seen its analytics investments turn into real business results. That has created some noticeable Big Data frustrations.

Research firm Gartner tracks market enthusiasm for emerging technologies with its “Hype Cycle,” and last year, Big Data moved from the “peak of inflated expectations” to the “trough of disillusionment.” While that sounds bad at first glance, it really means that businesses are moving beyond the stage of unrestrained expectations and instead starting to ask practical questions about how Big Data can actually solve their problems.

This more realistic view of Big Data means that when a project falls short of expectations, results-oriented executives may be less forgiving of the entire premise. But, is a lack of ROI an indictment on data analytics as a whole, or is it more a reflection of poor execution?

At Comptel, we argue it is the latter. As my colleague, Malla Poikela, wrote in a recent piece for LinkedIn Pulse, the most common hallmarks of a poor-performing Big Data initiative include difficulties accounting for every new raw data source and then turning all of that data into real-time contextual decisions and actions.

Successful programs rely on relevant actionability. Relevance comes from identifying contexts in real-time data, implying specific needs and employing predictive analytics to optimise target selection for those needs. Actionability is achieved through an end-to-end, integrated, real-time process that connects data streams through analysis to action.

It’s not about Big Data. It’s about Intelligent Fast Data, and it’s the only way to treat information at a time when technology empowers consumers to make informed buying decisions faster than ever and complexity grows in multiple dimensions simultaneously.

What are the benefits? With better understanding of existing customers and their preferences, operators can cue up the personalised service offers that customers want at exactly the right time on any device. It’s real-time marketing, driven by in-the-moment analysis, which leads to instant revenue opportunities.

More generally, Intelligent Fast Data can be considered a process that constantly monitors various forms of digital demand and connects that demand with available digital supply, be it a subscriber needing faster bandwidth temporarily to watch a video on demand, a network requiring additional capacity from virtualized packet core functions or supplying a service desk with a data feed from temperature sensors in a connected home.

Here’s how operators can start to make the switch from Big Data to Intelligent Fast Data.

Think Beyond Rules-Based Parameters

One of the downfalls of traditional decision-making system implementations has been a sole reliance on rules-based infrastructure. This form of analytics provides recommendations based on a set of pre-determined rules, but the challenge is that such a system might not be very accurate and can become overly tedious to manage as the number of rules increases. Rules or logics are important decision-making capabilities, but just like in human decision-making, they often need to be supplemented with capabilities such as pattern matching, predictions and anomaly detection. Intelligent Fast Data enables just that: the embedding of machine-learning-driven advanced analytics capabilities into decision-making.

If Insurance, a property and casualty insurer, took this approach to revamp its insurance claim analysis. If stepped up its automation capabilities with an Intelligent Fast Data system, which automatically learns patterns of insurance claims and flags normal claims for automatic processing, while highlighting potentially fraudulent or anomalous claims for further inspection. With the Intelligent Fast Data system the company was able to further reduce manual claims processing work and triple its number of accurately processed claims.

By embracing Intelligent Fast Data (i.e. decision-making automation with embedded analytics), digital and communication services providers can speed up and enhance the process that turns their data streams through analysis and targeted actions into new revenue streams.

Eliminate ‘Data Wrangling’

Another obstacle that could be holding back your switch to Intelligent Fast Data is a phenomenon known as “data wrangling.” According to the New York Times, data scientists can spend 50 to 80 percent of their time and talent essentially prepping data for the analytics process. It’s busywork, and it means you could be taking far too long to turn customer data into action.

To eliminate time-consuming data cleansing and enable faster time to action, a flexible and agile data processing layer is required, particularly one with the ability to integrate information from any digital source, then automatically cleanse, normalise, enrich and transform the data into ready data products and actions, which are consequently delivered to the systems with specific demands. Such a data processing layer must have smart adaption capabilities so that is able to cope with changes in data streams and the addition of new ones without data wrangling.

Remove Purchasing Friction

Changing how you integrate and process data is step one to drawing more value from Intelligent Fast Data investments. However, operators also need to eliminate any potential roadblocks to realising revenue from the insights data provides. This sometimes requires creative solutions.

For example, Indosat, one of Indonesia’s top mobile operators, needed to find a way to monetise mobile revenue opportunities in a country with one major roadblock. Despite being home to more than 250 million residents, only 8 million people in Indonesia have credit cards. That’s 3.3 percent of the population and 7.7 percent of the country’s sizeable base of smartphone users.

Smartphone users in Indonesia can’t simply purchase apps and services on their phone from a stored credit card like consumers elsewhere. However, a creative solution – direct carrier billing for the Google Play store – enabled Indosat to offer its consumers the same purchasing experience smartphone users worldwide enjoy.

Removing this obstacle opened up a new revenue channel for Indosat, and as the operator collects customer app usage data, it will be able to refine this information into insights and actions that drive even more financial benefits.

Intelligent Fast Data, ultimately, allows operators to profit from a wealth of Big Data.

Want to learn more about how Intelligent Fast Data can help you draw more value from new and existing customer relationships? Download our new book, Operation Nexterday, for expert research and insights.


Comptel and Big Data at Mobile World Congress 2014

Posted: February 24th, 2014 | Author: Matti Aksela | Filed under: Behind the Scenes, Events | Tags: , , , , , , | 1 Comment »

You probably have not been able to avoid hearing the term, “Big Data”, nor about the expectations of its limitless possibilities for communications service providers (CSPs). CSPs have a unique opportunity to delve into the spectrum of network, customer, service and other information at their fingertips and flowing through their OSS/BSS, eventually using it to improve both internal operations and customer-facing processes.

But Big Data sadly often means Big Projects. It is not just management of the three core “Vs” – volume, velocity and variety – that contributes to this, so can the setup of the technology to store and collect the data. But often, the biggest challenge stems from the fourth “V”, or value. What do operators need to do in order to drive true value from Big Data? I believe that there are some key requirements for being successful here:

  • Have a strategic business objective to focus on. Do not just collect data for the sake of collecting data, but have a goal in mind and a roadmap of what to do to drive more value when you reach that goal. (Buy-in from the boardroom, of course, helps, too, especially with issues like breaking down organisational silos.)
  • Don’t start with a blank slate. It’s important to have a set of proven, productised applications to address your business pains, whether it be customer experience-driven like smart throttling or network-focused for proactive service management, for instance.
  • Collect experience and learning in your organisation if you see information as your key asset, but don’t wait until you have built an experienced team to do so – have that as your plan, but start generating value from operational applications from the get-go.

And that’s where Comptel comes in. We’ve been developing our Big Data offering to help CSPs give their initiatives a running start, and also  supply them with the tools to support information-based decision-making and derive the true value they’ve been looking for – quickly and in a future-proof and extendable way – to solve acute business pains and build on that success.

Comptel provides a true Big Data solution, addressing all key components of the Big Data process:

1)     Data Ingestion: Integration, importing and formatting of historical and real-time data from CSPs’ own data sources, combined with external data for a truly holistic view of the business. This is powered by Comptel’s proven technology used in our mediation solutions.

2)     Data Management: Transformation, correlation, enrichment and manipulation of data to ensure optimal usability, and using the most appropriate methods to store data—whether it be Hadoop for unstructured data, massively parallel processing databases or in-memory data grids.

3)     Data Analysis: Highly accurate, real-time predictive analysis, modelling and reporting, powered by machine learning.

4)     Business Analytics Applications: Productised solutions to solve acute business pains, utilising the whole Big Data solution to drive immediate value.

One important aspect of Comptel approach’s is the utilisation of both historical and real-time data to drive true value—we do not see these as separate discrete steps of a process, i.e. building a predictive model and then applying triggers based on the model’s predictions, but instead having the predictive model applied in the real-time data stream to reap and benefit from contextual intelligence.

Clearly, Big Data analytics is reshaping the telco landscape. According to our recent research supported by Vanson Bourne, about two-thirds of telco executives (64 percent) say they are already in the process of leveraging Big Data to improve customer service, for instance. This is the time for Big Data to show that it is not just a hyped concept but a true generator of value—and we believe that the way to do that is through scalable, Big Data solutions designed to achieve and build on CSPs’ business objectives from day one.

To discuss how contextual and operational intelligence can augment CSPs’ efforts, come to our booth (Hall 5, Stand 5F41) at Mobile World Congress 2014. Look forward to seeing you in Barcelona this week!


Want to learn more about telco in 2014? Download our new eBook, “What Telco CMOs and CTOs/CIOs Are Thinking in 2014.”

In this eBook, we share exclusive, global executive research that highlights:

- Executive strategies for 2014

- Barriers to integration

- Technology priorities

- Attitudes toward data & planning

Download