Independently of its size and economic sector, most companies feel a certain pressure to embark on big data analytics projects. While we agree that data has become one of the most valuable assets of any organisation, from our perspective, having worked with operators and equipment vendors, we believe there are certain use cases that bring more value to telecom organisations than others.
Also, we believe certain considerations should be made before starting working and investing time and other resources in this area. Based on our experience from projects for network service operators and equipment manufacturers, we have gathered some advice that most organisations may employ with great benefit, and that we would like to share here.
So if your company handles big data sets in any form, keep on reading. Sooner or later you will need to extract your data’s real value in order to gain efficiencies in operations, optimise performance or improve customer service.
Any successful project starts with precisely defining objectives and justifying it from a business point of view. Apart from having a fair understanding of advanced analytics techniques and what they might achieve, big data analytics projects in any organisation also require a realistic planning and resource definition. We therefore suggest analysing the following questions in the planning phase:
- What are the objective(s) with the project? If various, prioritise clearly.
- Who controls the data in the organisation?
- Where and in what formats are the data stored and for how long? Will they be easily accessible within a reasonable amount of time?
- Are the necessary skills and expertise already in place in the organisation, or is internal training or external expertise needed?
- In what specific areas will an analysis of historical data be most relevant and in which ones will real-time analysis have to be executed? Real-time or near real-time analysis is not always indispensable, in some cases historical analysis will have a higher impact on objectives, as well as often being less complex.
- How will the results of the project be employed further in the organisation? For example, are the right teams trained and ready to work in a predictive manner? The right mind-set is as important as skills if new ways of working is to be implemented successfully.
- Is the project justified from a business point of view? If not, refocus your objectives.
Minimum Viable Projects
In Blue Telecom Consulting we have developed a successful formula that includes measuring the grade of maturity of our customers when it comes to big data analytics before engaging in a project. We evaluate the state of their security and data access policies, the skills sets in place and human resources available for the project, among other variables.
If the right conditions are met we then define partial objectives for which we design smaller projects. And of these, only a couple of the most promising ones with manageable dimensions are initiated. In a way we follow the same logic that lean start-ups do when launching “minimum viable products”, rather than delaying their launch until a more finished product is ready. We follow this approach since partial results may be achieved very rapidly and with limited initial resources, which gives a sound basis for continue working and refining. We believe this is better than starting out more broadly or ambitious. Our formula is also highly motivating for the teams involved and allows bringing quick, visible results to management.
Most Common Operator Use Cases
Our experience has shown that for network operators there are three use cases for which big data analytics projects are most often legitimised:
- Prediction of alarms of network events, root cause analysis and the automatic resolution of these
- Predictive network planning and optimisation
- End-to-end monitoring of network quality
These use cases, tailored to the operator’s specific environment, enable monitoring the quality of service offered to customers, the anticipation of events and optimisation of networks. Thus, the operator enhances the value of its services, which in itself justifies investing in big data analytics.
And even if this step-by step way of working might seem a little tedious, we consider this to be optimal for telco organisations given their technical complexity and other characteristics. The initial preparative phases are of crucial importance and decisive to whether introducing big data analytics will be a success on which to build or not.
Josep María Balaguer