TrafficGuard: Standing in the way of ad fraud to help businesses reach more customers
Greatly reduced the time to build batch data pipelines
Faster time to market for advanced analytical needs
Reduced the burden of infrastructure management and maintenance
TrafficGuard is a global advertising verification company, helping advertisers to avoid wasted spend and unlock the best outcomes. Powered by the immense scale of data and the expertise behind their universal advertising verification suite, TrafficGuard drives superior search advertising performance by verifying advertising engagement as they occur, proactively blocking invalid traffic from infiltrating search campaigns; helping ad spend to reach more real users and protecting the integrity of data that marketers, agencies, designers and developers rely on day in, day out to drive performance improvements.
TrafficGuard is based in Australia, headquartered in Perth. The company also has offices in Singapore, and has staff presence also across the UK, Brazil, India and the US, serving over 3000 customers globally.
TrafficGuard was a happy customer of AWS for many years, but as their business needs evolved and their infrastructure requirements became more elaborate to accommodate ML, they found more alignment with Google Cloud.
“As we started to look into more elaborate uses of the cloud like running analytics and machine learning (ML), we found that the alignment with Google Cloud was far better than our previous cloud provider, with Google being predominantly an ad business,” says Luke.
Kasna executed a multi-phased discovery, architecture and migration approach to re-platforming to Google Cloud. This included assessment, tech-spikes and prototyping of required technologies to ensure functionality, performance and scalability requirements were met. New services were set up and deployed to TrafficGuard’s Google Cloud environment via fully managed and serveless components. New data pipelines were created and existing ones modernised using Dataflow. Applications were migrated and modernised using Cloud Functions and Cloud Run.
By taking away the need for server infrastructure maintenance, Google Cloud enables the business to focus more on its ad tech solutions and verification. It also means that TrafficGuard can keep a lean team to manage its proprietary services.
“We’re all Engineers and Data Scientists, so we’re familiar with these platforms and anyone at the company can quickly spin up environments to get access to the data they need for testing,” says Luke. “Without having to worry about infrastructure, rapidly build prototypes to continuously iterate, innovate, and find better solutions for our customers.”
Working with Kasna has enabled a smooth migration to Google Cloud and helped TrafficGuard benefit from the expert cloud product knowledge of specialists.
“Kasna has been a great partner. They took the time to really understand our business, are process-driven, and very transparent in what they do so it has been easy to work with them,” says Luke.
Key Products/Services Used
- AI Platform
- Cloud AutoML
- Cloud Bigtable
- Cloud Composer
- Cloud Functions
- Cloud Run
- Cloud Storage
- Cloud Tasks
- Compute Engine
- Cloud Dataproc
- Cloud Memorystore
- Cloud Pub/Sub
- Google Data Studio
One of the most compelling factors for migration was BigQuery, with which the team immediately found tremendous value in. BigQuery can process information quickly, even for large datasets, and allows for streaming inserts, and lowering the total cost of ownership of a real-time pipeline.
We were able to rapidly build a custom time series reporting engine for external facing products that can handle large volumes of data across large time periods.
This was a great fit for our type of data. BigQuery supports advertising data characteristics; high dimensionality, cardinality and scale without issues on cost, query, and operational complexity. The cherry on the cake is that the team can perform advanced analytics with SQL.
For serverless compute requirements Cloud Run and Cloud functions provide TrafficGuard tremendous value in operational efficiency with the pay-per-use model. All of this is monitored and operated using Cloud monitoring.
Being able to process streaming data and then use BigQuery for compute, storage and ML simplifies a lot of TrafficGuard’s processes.