Mobile network operators (MNOs) have the highest degree of transaction-intensity of any comparable service provider sector. They process millions of voice calling and data messaging records on a daily basis, in addition to massive streams of related information i.e. subscriber service orders, cellular network reports, customer care inquiries and customer billing data. Moreover, by carrying traffic from social media applications and ‘over the top’ instant messaging utilities, MNOs now have unique access to the personal and business-oriented interactions that their subscribers generate via their use of wireless-enabled services. Increasingly, operators are turning their attention to the strategic and operational implications of this market phenomenon and in particular how they align their organizations to address this ‘Big Data’ opportunity.
What is the starting point for operators?
Unlocking the potential of these captive information assets – structured and unstructured, internal and external – demands that MNOs have a coherent and integrated data strategy.
The principal elements of any such strategic template include:
- Establishing an enterprise-wide ownership framework for the end-to-end management of data – from the moment of creation through to the point of use.
- Setting clear goals for how the data will be used – whether for optimizing the service delivery platform, developing closer relationships with customers and anticipating or responding to market changes.
- Determining the detailed policies and procedures that govern how data is collected, analyzed, and acted upon.
How best are operator data assets managed?
The ‘D4’ Process Model is a standard framework being adopted widely within the global network operator community. It characterizes the four primary development phases of data asset management in complex service delivery environments.
Each of the four phases has its own inherent complexity, and rigorous discipline is required in order to maintain the integrity of the end-to-end data management process.
- Data acquisition. Quality-tested, contextually-relevant and easily interpreted data is a desirable pre-requisite for downstream analysis and action. However, this level of quality requires significant workflow planning and investment in systems and processes – even machine-generated and curated data is error-prone.
- Discovery. This is the phase of the process where most management attention is currently being focused. The fierce competition for data sciences talent is a direct consequence of the need for companies to create first mover in-house analytics capabilities. And the various insider debates continue to rage i.e. “does math now trump science?” and “will intuition always beat big data?”
- Delivery. Presenting the final analysis results to a key decision-maker and securing their trust in and acceptance of the findings is the “acid test” of any Big Data initiative. Enhancing an existing business process such as new service development with fresh data-driven insights is another benchmark of credibility. The investment in data visualization tools is evidence of the criticality of this step.
- Dollars. Profitable revenue generation or measurable business value creation is the end game of any Big Data initiative – ROI is the only metric that counts ultimately.
The aggregated customer information being accumulated on a daily basis by MNOs holds the promise of augmenting the existing datasets available to third party organizations – i.e. public and private service organizations, advertising agencies, media publishers, government agencies, etc.
This is the potential pay-off for the MNOs in respect of their investment in implementing these end-to-end data management processes and ultimately to building out a value-producing customer data organization.
The D4 Process Model provides a mechanism for embedding structure into the planning and implementation of early stage Big Data initiatives.
What are the preferred organization models?
MNOs must also define the right organizational structure for managing any significant Big Data project – one that can effectively align the demands of the business with the technology requirements needed to support those demands. Such initiatives necessarily involve an element of cultural change within the operator, as they require collaboration across conventional functional and business unit silos.
Arriving at the appropriate cross-functional, cooperative model for leveraging data assets involves some degree of compromise and trade-offs between the various stakeholder groups. Operators begin at different points along the evolution path, with different organizational structures and levels of maturity, and varying sets of capabilities. There is no ‘cookie cutter’ approach at this stage of the Big Data life cycle.
There are three discrete organization models that are currently being implemented across those MNOs who are actively pursuing Big Data initiatives:
- IT systems-led (including a dedicated data analytics group)
- Business function-led (with Marketing and/or Finance as the lead functions)
- Matrix organization (a hybrid IT-Marketing-Finance group)
Given the clear potential for creating new and sustainable revenue streams from Big Data initiatives, it is likely that the majority of MNOs will elect to form matrix organizations over time as a critical mass of Big Data resources is established.
This structure implies the leadership of senior executive who understands the MNO business and has a clear line of sight into C-suite priorities. In addition, this leader must have considerable authority to make key decisions about investment priorities with regard to both IT and business resources, a high degree of respect among both business and IT stakeholders, and the ability to balance short-term business needs with longer-term goals for building out the necessary data capabilities.
As such identifying and attracting such talent (if unavailable internally) is a potential drag on progress in gearing up to attack the Big Data opportunity.
The war for talent is not confined to the fierce competition for data scientists!