Delivering faster, more granular insights for clients following a switch to Google Cloud
Google Cloud results
- Halves batch processing times to make up-to-date data available faster
- Powers on-demand reports for clients in seconds with BigQuery
- Delivers a more flexible, more reliable infrastructure with Google Kubernetes Engine automatic scaling and auto repair
Managed services reduce DevOps load by 40%
Brands that spend big money on digital campaigns want to know that their messaging reaches the right audiences. But in a complex digital landscape, that isn’t always easy to achieve. According to Forbes, the issues of fraudulent traffic, a lack of transparency in delivery and reporting, and the ability to track real ROI were among the top challenges facing digital advertisers in 2021.
That’s where Adloox comes in. Its all-in-one platform helps advertisers and media agencies to address these issues, enabling them to make sure ads are viewable, impactful, and delivered in a brand-safe environment. It measures every supplier, domain, ad placement, and ad unit and then provides objective recommendations to help maximize the performance of campaigns. The platform also detects and filters out fraudulent traffic, and makes sure ads don’t appear alongside inappropriate content, such as fake news, by analyzing multiple page elements in real time.
To deliver these insights and benefits for its clients around the world, Adloox processes huge amounts of data, with more than 7 billion fraudulent bid requests avoided every day. When it launched in 2009, it built its application infrastructure and data pipelines using dedicated servers. But as the business grew, it realized that switching to a cloud-based architecture would enable it to automate its processes, freeing its team to focus on development. With the support of implementation partner OP-Rate, Adloox migrated to Google Cloud, gaining the scalability and improved performance it needs to deliver insights to its growing client base.
« The digital advertising sector has been transformed over the past few years, because of the increased automation of auctions and bids, » explains Romain Bellion, CEO and co-founder at Adloox. « To deliver the best service for our clients, we need to be able to process very large volumes of data, and manage those volumes in ways that enable us to generate precise reports as quickly as possible. »
Supporting growth with managed services
With its previous infrastructure, Adloox had to provision and deprovision additional machines in advance. It wanted to improve the scalability of its infrastructure to handle large peaks in traffic and to have the flexibility to scale down dynamically, as its traffic is often three or four times higher during the day than the night. This meant it was constantly over-provisioning and monitoring its infrastructure to make adjustments.
Adloox wanted a solution that offered managed services, so it could concentrate its efforts on developing new application features and fraud detection algorithms. Security was also an important concern, as keeping systems up to date was also time consuming. « We looked at other solutions, but as our platform offers native integrations with Google advertising solutions Google Ad Manager, Display, and Video 360, we already had a trusting relationship with Google, » says Bellion. « We knew we would be well supported during the transition, so it made sense to extend our partnership to Google Cloud. »
Keeping reporting live during the migration
To ensure it would be able to contixnue to serve its clients’ live campaigns during the migration, Adloox chose to work with OP-Rate. « We took six months to find the right implementation partner and the right approach to make sure the process went well, » says Bellion. « We chose OP-Rate because of its previous experience within the adtech sector, which meant that it immediately understood our needs. It provided very precise advice on how to manage the migration, and which services to use to build a functioning architecture. »
Together with OP-Rate, Adloox built a parallel architecture and moved its services over load by load, ensuring each load was functioning well in production before moving on to the next. The first stage involved moving its existing data platform over to Google Cloud, optimizing its dataflow pipeline in the process. Its data processing services are now orchestrated with Google Kubernetes Engine (GKE), using Cloud Load Balancing to deploy in three regions and Compute Engine for additional processing. During the day, data is stored using Cloud Storage, and overnight Adloox uses Dataproc to run its fraud detection algorithms and other calculations. The outputs are then stored back in Cloud Storage, or fed into BigQuery to run client reports, accessed by clients in the Adloox dashboard.
Next, its tracking applications were moved over to GKE. « Switching to Google Kubernetes Engine has made our system more resilient; if a machine fails, then it’s replaced automatically with another. That’s really important for us, in terms of making sure we can deliver to our customers, » says Bellion. Adloox also uses Cloud Logging and Cloud Monitoring to monitor its infrastructure both through retrospective logs and real-time alerts, as well as Cloud SQL for its relational database needs and Cloud Composer and Cloud Scheduler to automate workflows.
Offering faster, more granular reports
Now, if Adloox experiences increases in the volumes of traffic on its platform when traffic might be three or four times higher than usual, its infrastructure automatically scales to meet the demand. Since migrating to Google Cloud, it’s also halved the amount of time it takes to complete its overnight batch processes. « Not only is our batch data ready twice as quickly, we’ve also been able to reduce our costs and have eliminated around 40% of time spent on infrastructure management, » says Bruno Galeppe, co-founder and CTO at Adloox. « That means we can focus on developing new features, which has helped us launch the new version of our platform faster, for example. »
Switching to BigQuery means that Adloox can provide clients with new and enhanced analysis tools too. « Powered by BigQuery, clients are now able to create reports on our platform based on three, six, or twelve months of data in under ten seconds, » says Bellion. « Our infrastructure on Google Cloud enables us to deliver insights that previously we wouldn’t have had the processing power to calculate, and it’s ready almost immediately. »
As well as further expanding its usage of Google Cloud-managed services, Adloox is looking into the possibility of using Vertex AI within its R&D department, to enhance its fraud detection algorithms. It’s also looking forward to launching its platform and client dashboard based on Google Cloud for all its clients in March 2022, as well as new fraud detection functionalities.
« Google Cloud is a natural fit for a business like ours working in the Adtech sector, » concludes Bellion. « As well as accessing best-in-class products, the support we’ve received from the Google team and from OP-Rate has been an important part of the success of our migration. »
Adloox is an all-in-one ad-verification and insights platform that offers transparent and granular data, helping advertisers and media agencies to make the most of their digital media spend.
- Google Cloud
- Kubernetes Engine