Top UK online betting platform cuts gambling fraud by 84% using Elastic Observability

In addition to managing system capacity, the Infrastructure Platforms team must also support the betting company’s efforts to prevent fraudulent activities such as identity theft and money laundering. Their approach is based on spotting patterns of gambling behavior, which also enables them to intervene on what is deemed risky customer activities.

Data gathering, enrichment, and analysis are essential for the success of these strategies and the overall health of the betting platform. This is where the team behind Infrastructure Platforms play a vital role. They deliver shared services to technical teams across the business, including enterprise secrets management, messaging services, configuration management, and CI/CD (continuous integration and continuous delivery) tooling. The team also provides logging services that monitor a range of activities from system health to user behaviors.

Until around 2018, the team used a logging solution that wasn’t flexible enough to manage fast-growing, uneven log volumes. “The license for our previous logging software was based on a maximum amount of ingest per day. This wasn’t suitable for us where we can have several quiet days, followed by a very busy period as the result of a high-profile soccer game. We wanted a more adaptable solution to cater for these events and allocate resources to pipelines on busy days - the flexibility of Elastic was the answer,” says the engineering manager responsible for infrastructure platforms.

When searching for an alternative solution, he consulted with his colleagues, some of whom had previously worked with the Elastic Observability logging solution in other roles.

In order to ensure integration and deployment was managed effectively, the betting company’s team utilized Elastic Consulting services to help them seamlessly migrate data from the previous solution into Elastic.

Today, Elastic Observability handles all major logs including application health and gambling activity patterns. Logs are shipped to Elasticsearch for enrichment using Logstash. They are then shipped to Elasticsearch clusters running in Elastic Cloud Enterprise. This enables the Infrastructure Platforms team to maintain separation between internal customers using independent clusters and separately hosted Logstash deployments in a Kubernetes container-hosted environment.

The Infrastructure Platforms team also took advantage of Kibana integration with Slack to trigger alerts to the relevant monitoring team. Each Slack message contains a link that opens the relevant Kibana page so that a team member can track the issue and intervene quickly.

During the migration to Elastic Observability, the Infrastructure Platforms team spoke to many internal customers to better understand how the new technology could support their objectives. Marshall highlights the importance of helping the Fraud and Risk team to identify illegal activity.

With Elastic Observability, Fraud and Risk can search datastores across multiple systems to detect fraudulent accounts signing up, to take advantage of free bets and bonus offers. This information is analyzed by artificial intelligence software and suspicious activity can be stopped instantly. If there is a need for human intervention, the system creates an alert directly in Slack so that a specialist can review the activity and respond as needed.

Additionally, Elastic Observability enables the business to adhere to regulations for safer gambling. “We can monitor patterns of deposits and withdrawals to identify at-risk customers and even intervene as the bet is placed if needed,” says the company’s Automation Engineer.

In recent years, the firm has introduced safer gambling tools including deposit limits that enable customers to set their budget, and self-exclusion where they can block their accounts from six months to five years. This applies to all of their sister companies, so if someone excludes themselves from one betting product, they are also blocked at the other brands across the group which consists of up to five other betting platforms. “With Elastic we can easily share back-end data across all our brands so that the customer is always protected, when they are betting using another platform,” says the company’s automation engineer.

Importantly, and a major benefit to the company, is that they are using Elastic Observability’s flexible licensing model to reduce costs and manage budgets more effectively.

The business also appreciates that it’s inexpensive to try new Elastic features. “There's very little cost when adopting a feature on a small scale and using it to demonstrate value to the business. This makes it easier to secure funding for further licensing based on this evidence,” he says.

Although the Infrastructure Platforms team is largely self-supporting, the company welcomes the opportunity to call on Elastic when de-snagging the deployment of new features.

"With our safer gambling measures, we did have an issue with data volumes creating too many buckets in an aggregation. We reached out to Elastic, and they gave us a really good solution involving data transformations. This is something we can use in the future if we encounter issues around aggregation volumes," they said.

Additionally, the product team who own the Elastic solution, also stresses the importance of Elastic Professional Services who assisted the company with the platform migration from the legacy solution. “It was a large undertaking, so the assistance of Elastic Professional Services was invaluable for finding solutions to technical issues,” they say.

Looking to the future, the team are also testing Elastic APM to pinpoint and fix application performance issues quickly, while continuing to set an example in the prevention of fraud and the protection of vulnerable customers.