Big Data is the biggest buzzword today in the industry. Almost every data centric company is trying to figure out what Big Data means for them and how it can help them do better. The Boston Consulting Group does a fantastic job in this article to summarize Big Data and highlight key aspects.
After reading-up on the topic and having insightful discussions with number of thought leaders, I feel that following are the key steps that a Mobile Network Operator needs to take before embarking on the Big Data journey.
- Identify all key data sources
The first step is to identify all key data sources that a Mobile Network Operator uses to make decisions. Typical data sources for a MNO include but are not limited to voice and data usage, billing information, CRM data, call center records, network data, sales channel data, financial management system, web usage, digital and social media, and data from market research.
- Separate ‘structured’ data sources from the ‘unstructured’ ones
From the data sources identified in first step, the next step is to identify data sources that are ‘structured’. Structured data sources are usually available in analyzable form. The enterprise Data Warehouse would ideally consolidate all structured data sources such as Usage, Billing (CDRs), CRM data, and Sales data among others.
- Convert ‘unstructured’ data into ‘structured’ data
The reality is that there will always be data sources that generate huge volumes of unstructured data. In case of a typical MNO, it is likely that data sources such as network data of customers (calls drops, data speeds, voice quality, geographical information etc.), web usage (websites visited, type of applications used etc.), call center records, and social media conversations contribute to huge volume of unstructured data. It should be part of the Big Data strategy to convert such unstructured data sources into structured data sources with help of various initiatives. For example Service Quality Management (SQM) and Network Customer Experience Management (NCEM) are two initiatives that aim to collect and structuralize the huge network data. Similarly, Deep Packet Inspection (DPI) methodology/technology can be used to analyze the unstructured web usage of customers. There are elaborate Voice of the Customer programs and software available today that analyze social media conversations and call center records.
- Have a vision on what to achieve from Big Data
Once it is clear what data sources are characterized as unstructured, the next step is to crystalize a vision of what objectives the organization wants to achieve by analyzing the big unstructured data. The outcome of the vision and objective setting exercise would be a set of initiatives with specific goals that Big Data should achieve.
- Use data mining to generate insights from unstructured data
While initiatives to convert unstructured data into structured data run in parallel, one of the most direct application of Big Data projects is to use advanced algorithms and data mining techniques in making sense of unstructured data to generate meaningful insights for decision making that are aligned with the vision and objectives of big Data. Data mining techniques are not restricted to be used only for unstructured data. In-fact it works even better on structured but complex data structures. For MNOs one key area that data mining can help generate actionable insights is Churn Prediction Modeling.
- Work towards creating ‘Single Customer View’
The end goal of Big Data journey for any organization is to make better-informed decisions by knowing more about its customers and to eventually increase profitability. The Big Data journey outlined in steps above should help organizations achieve this objective by linking all of this to create a Single Customer View from multiple structured and unstructured data sources. For example in case of an MNO, the Big Data journey should help the company in knowing each customer as a person (CRM data) with specific usage behavior (CDR/Billing data), personal preferences (web usage/social media data), emotional relationship with the company (call records), and information on the Quality of Service (Network data) experienced by the customer. Data mining works on this aggregated information to predict usage, churn, and behavioral propensities. Such information helps marketers and business managers to make informed decisions that improve Customer Experience, enhance customer loyalty, and positively impact the bottom line.
Looking forward to your comments and feedback. Please note that these thoughts are strongly influenced by my personal experiences with Mobile Network Operators. Your experiences would most likely be different and unique from mine. I would love to know about your experiences and thoughts about Big Data and how is your company going about it.