Giffgaff goes AWS to drive operational transformation

UK mobile virtual network operator (MVNO) Giffgaff, which runs on the O2 network, has selected Amazon Web Services (AWS) to base all its technology infrastructure and application development, becoming the first MVNO in Europe to go all-in on the AWS cloud.

In the deployment, Giffgaff will use more than 60 AWS services, including compute, analytics, storage, databases, containers and machine learning. The company says it will aim to analyse and understand better in real time which services its members are using, what new services they want and then quickly enhance and introduce new services. By the end of 2020, it will migrate its remaining applications and databases from its existing on-premise datacentre to AWS.

By using AWS, Giffgaff said it has evolved its development lifecycle from a complex, monolithic approach, with limited experimentation and which slowed the business down and ultimately hindered innovation, into a modern, microservices-based architecture that has enabled rapid development and deployment cycles, and empowered its teams to increase innovation.

Before moving to AWS, it could take Giffgaff up to two weeks to provision a server; now the operator can perform this task in minutes.

Giffgaff said its data scientists use AWS analytics and machine learning services to better understand its members’ network experience and anticipate future needs. When members report an issue, Giffgaff can analyse the text of the report, classify the issue, and direct it to the appropriate agent for resolution.

By aggregating this data across all incoming cases and then analysing it, the company can create an early warning system for network incidents. Giffgaff uses Amazon Kinesis to collect, process and analyse member data in real time to react quickly and increase member satisfaction and engagement by delivering personalised marketing and member care.

Giffgaff also uses Amazon Elastic Compute Cloud (Amazon EC2) Spot instances to enable customers to take advantage of unused Amazon EC2 capacity in the AWS cloud to minimise cost. Spot instances are available at up to a 90% discount compared with on-demand prices.

By adopting a continuous delivery approach and moving its containerised workloads to a fully managed Amazon Elastic Kubernetes service, Giffgaff is calculated to have freed up 3,000 days of engineer and developer time to allow them to concentrate on creating and introducing new applications. That, says AWS, is the equivalent of refocusing 14-15 people on innovation for members, instead of the undifferentiated heavy lifting of managing servers.

Also, by using AWS, Giffgaff has increased the pace of feature releases to more than 9,000 times a year, compared with just 10-12 times a year when using on-premise infrastructure. This has helped drive an increase in Giffgaff membership.

“We started out with a traditional, on-premise infrastructure, but the need for ongoing maintenance made this model overwhelming for our technical team,” said Steve MacDonald, chief operating and technical officer at Giffgaff. “For example, it used to take us up to two weeks to provision a new server. When we began to adopt AWS, we were able to turbocharge our development lifecycle by focusing on innovation rather than wasting time on maintenance.

“It’s such a powerful capability for a digital-native business like ours. The process of constantly evolving our software means we can always deliver an award-winning service to our members. Our plan now is to extend the use of AWS and make our members’ voices serve as one of the primary data sources for business decision-making.”