GCP – Modernizing telecommunications and keeping the UK connected

When Virgin Media and O2 joined forces in June 2021, we made it our mission to modernize the country’s connectivity. By the end of that year, we had completed upgrading our fixed broadband network with download speeds of at least one gigabit per second, supporting the UK government’s drive to take gigabit broadband nationwide by 2030. Gigabit downloads would allow a high-definition film to be downloaded in under a minute.

As we continue with our rollout of our 5G mobile networks, we are not only simplifying connectivity for our 47 million customers, but helping to revitalize the UK economy by connecting individuals, businesses and communities up and down the country. 

Modernizing a nation’s telecommunications means we need to be a modern, data-driven company. But the tech stacks of both Virgin Media and O2 were built on a complicated patchwork of legacy systems inherited from companies they had absorbed throughout their history. So when the time came to merge the two businesses’ architecture, it was clear that to integrate our systems and get the most from our vast data pools, we needed to do so in the cloud.

Our migration to Google Cloud gave us a fresh start to use its intuitive solutions which suited our needs to build a new infrastructure in a fast and cost-efficient way. At the heart of that new build was our data architecture, which would enable us to transform our organization from a traditional telecommunications business to the 21st-century tech company our customers demanded. 

The challenge of data analysis without data

As data analysts, our goal is to use the data at our disposal to gather insights to increase efficiency, target customers and improve our products. However, we don’t always have a precise question in mind from the outset. We come to the data with high-level problems we want to explore, we ask a question, get an answer, and then refine our journey from there. Iteration is the key. 

Before we migrated to Google Cloud, iteration was impossible. Each query took hours to process, which meant that refining a simple question would have taken days, or even weeks, assuming we hadn’t lost our train of thought along the way. Add to that the fact that our overnight jobs began at 6pm, effectively killing off all open queries, and it’s clear how limited we were in being able to take data-driven decisions. We were unable to segment customers to personalize our offering. In the event of outages on the broadband network, it took us hours to understand what the problem was, and then a six-month lead time to correct any algorithms in our modeling. We realized that if we wanted to revolutionize the telecommunications industry, we would first have to revolutionize our own business. 

Illuminating a path to powerful insights with Google Cloud

Migrating our data to Google Cloud has been like switching on the lights in a dark room. Now when we want to know something about our business, with some of the  query execution time reduced from 16 hours to 40 seconds, we can refine our way to valuable insights within hours, if not minutes, instead of days or weeks. What’s more, the average time for manual testing is now four minutes, as opposed to 44 days, while we experience almost no monthly downtime for our data analytics applications, as opposed to the two hours a month we were used to. And because Google Cloud feels like a tool that has been built for analysts, it allows us to use the data in a more agile manner. Working together in the cloud, our analysts and technical teams are able to react quickly to our business needs, and get powerful insights into the hands of our commercial teams when they need them. 

With most of our data from the broadband and mobile operations now unified in Google Cloud, we are able to build rich profiles of our customers, showing us how to target new customers and personalize our product offering to existing ones. And with data unified across domains, we are able to understand how a customer’s experience with, for example, their broadband service, might affect, say, their decision to renew their contract.  

In terms of our network monitoring, we have reduced the lead time for changes to our network model algorithms from six months to 60 minutes. This allows us to continually model any changes to the network before implementing them. Our network monitoring is now widely recognised as being far ahead of the industry standard, and it’s all thanks to our ability to use our data in Google Cloud to iterate at speed. 

Preventing carbon emissions with BigQuery and VertexAI

Google Cloud has also enabled us to use our data to reduce our carbon emissions. Every time our network engineers are needlessly sent out to fix a fault in the network, Pub/Sub ingests both the network data and the engineer’s analysis into BigQuery. With the help of VertexAI and Looker we are then able to model that data to understand and eliminate these unnecessary callouts, reducing our carbon emissions and saving us hundreds of thousands of pounds per year in wasted journeys.

A modern business, for 21st-century connectivity

The next phase of our data transformation is our data democratization program, which aims to put all this valuable data directly into the hands of the business departments that need it, enabling them to self-serve insights quickly and easily to empower informed business decisions. This is an essential part of our journey to becoming the modern telecommunications company that we want to be, where data underpins every decision we make.

As our migration journey nears completion, we have already experienced a major transformation in the way we operate, but we expect the next year to be even more impactful for both the business and our customers. One project in the pipeline is our plan to incorporate large language models into our customer service chatbots by leveraging all our existing data from previous customer interactions. Again, this would previously have been impossible, but with Google Cloud, we will be able to use this data to greatly improve the customer experience. 

As data continues to transform our business, we are excited by the ways we will be able to keep modernizing the UK, and continue connecting our customers long into the future.