Improving network spend optimisation involves a CSP understanding their capacity, usage and predicting trends and needs in real-time...
In my last post I made a big statement, "Customer churn is one of the biggest challenges facing CSPs today."
To address this issue, network optimisation is paramount to success. Improving network spend optimisation involves a CSP understanding their capacity, usage and predicting trends and needs in real-time.
There is a heavy reliance on the Business Intelligence department to provide accurate information, but the issue (as always) is time. The collaboration in many CSPs is there, but the data is so vast that making sense of it can take months, and during that period, the market has moved on…
Getting the right level of detail
Getting the right level of detail is challenging, it’s hard to convince the business of the value of making investments in a certain area, when you cannot explain the reasons behind your models.
Many CSPs – those that actually have big data solutions – spend much of their time and money working with general purpose data analytics suites to get basic answers to questions. Even then, a lack of data sources can produce results that are too summarised, which makes the business unable to act with total confidence. Across the industry, network planners and data scientists are being turned into programmers, often with limited business value to show for it.
If you don’t have a solution at all, the situation worsens… Being unaware of the detail behind your customers and their value leaves CSP’s open to the risk of ‘falling behind the curve’.
Consumer demand for high data consumption services is outstripping network capacity
Serious considerations for any Network Director are questions such as:
Are the peak over-capacity periods dominant?
What is my view of consumer service interruption as a % of use?
What OTT services used by consumers are causing the problem?
How do I decide the capex allocation to support the demand?
What is the blend of type and consumer network use that is growing?
What are the trends in network expenditure against revenue growth?
To answer these types of questions you are going to need analytics on your network usage and consumer behaviours. There are several different approaches to analytics that can be used.
Descriptive Analytics - the most common of these approaches. These technologies seek to categorise, explore, join together and summarise wide sources of data. Generally these can be thought of as exploring the “what is” of data.
Predictive Analytics - that seek to not only deal with existing data, but to calculate trends, patterns, movements and relationships and project this into the future to establish new data points. Generally these can be thought of as exploring the “what will be” of data.
Prescriptive Analytics - the final category and generally the most advanced. These take the patterns learned in predictive analytics and combine this with optimisation, simulation, modelling and regression to evaluate decision making and optimum choices. Prescriptive analytics provide insights to help answer business critical questions - for instance, “what is the best pricing model for this new handset?” Generally these can be thought of as exploring the “what could be” of data.
Using Prescriptive analytics solutions you could identify what the demand for data and network capacity looked like by content type, consumer profile, individual customer service experience or time and then related this to external, geographical and time sensitive events. … finally providing some answers to those key questions.