Media & Telecoms

Media & Telecoms

How cutting-edge analytics predicted network disruption and saved up to £1M in penalty fees

Our client manages a network of over 1000 radio and television transmission sites that supports broadcasters and mobile phone operators.

They suffered from network disruptions which would result in fines sometimes in the millions for short periods of downtime.

Their challenge was to identify which nodes needed maintenance and when, so that their engineering team could resolve potential network problems in advance.

Our task was to provide a tool that gave them early indications where there would be severe impacts to their network.

The highly profitable solution was delivered using an agile methodology:

  1. Analysis of current data issues and requirements
  2. Vendor selection – selected vendor with predictive analytics capabilities
  3.  Data discovery – understood what data was required from each system
  4. Data architecture – set-up ETL capabilities and transformations to bring data into a cloud data store ready to be consumed by the Analytics tools
  5. Implementation of Predictive engine and visualisation tool – crunched the data through a set of algorithms to predict which network nodes needed maintenance and visualised the results on a set of dashboards

Benefits:

  • Saved money on penalty fees (approx. £1m per disruption) – by predicting the network nodes that required proactive maintenance, they avoided potential future revenue losses brought about by unplanned outages as a result of equipment failure.
  • Enabled proactive maintenance – Instead of reacting to equipment malfunctions and failures that caught them off guard, the cutting-edge advanced ‘early warning’ system predicted which networks would go down, which guided and empowered engineers to proactively maintain specific issues reducing the risk of downtime
  • Made timely decisions – business users empowered to make timely decisions they needed to phase equipment out of the field responsibly making wiser procurement decisions.