OPTIMIZING ITS AUDIO ADVERTISING PLATFORM WITH GOOGLE CLOUD
Discover How to Scale Globally With Anthos and Google Kubernetes Engine
Specialized in the integration of audio ads into digital streams, SoundCast has built up its services using Google Cloud and Kubernetes. This French startup relies on Google Cloud's global network to be able to grow internationally and continue to double its revenue each year.
GOOGLE CLOUD RESULTS
- Doubling of revenue every year
- More than a billion invocations on the platform in one month
- Expansion on a global scale thanks to Anthos MCI
- One load balancer for multiple clusters in different regions around the world thanks to Anthos MCI
SOUNDCAST GREW INTERNATIONALLY AND CONTINUES TO DOUBLE ITS REVENUE EACH YEAR WITH GOOGLE CLOUD
SoundCast publishes a suite of solutions that simplify access to programmatic audio inventories and campaign implementation. It specializes in the integration of targeted audio ads across all types of digital audio streams (podcasts, web radio, voice-overs) on all platforms (mobile, desktop, tablet, voice assistant). The company was founded in 2018. From the very beginning, Google Cloud has been essential to the development of this unique and innovative platform. Already highly successful in France, SoundCast is now focused on expanding internationally, once again with the help of Google Cloud solutions.
EXTREMELY HIGH DEMANDS FOR PERFORMANCE AND AVAILABILITY
SoundCast is so much more than just a technology partner—it also supports its customers in their business marketing strategies. SoundCast’s solutions are connected to all automated sales platforms in real time, allowing publishers to sell their audio inventories (podcasts, live radio, web radio, streaming platforms) to media agencies and on ad-exchange programmatics. « We’re more than just intermediaries between advertisers and publishers, because our platform delivers tremendous value, » explains Patrice Sguerzi, CEO of SoundCast. « Really, we act as intermediaries, facilitating bidding transactions that happen in 200 milliseconds, but where several parameters must be taken into account to make the right decision. »
The SoundCast platform processes dozens of criteria in real time for every new request, including age, gender, region, or even the file format accepted by the target platform and the placement of the advertisement (pre-roll, etc.). Simultaneously, all broadcasts are tracked so customers can see the results of their campaigns in real time.
Put simply, the SoundCast platform has extremely high standards when it comes to performance and reliability, but is also subject to very high load variability. The number of requests processed in real time depends not only on the scale of the campaign being implemented at that given moment (number of users listening to a broadcast), but also on the number of campaigns being managed simultaneously, whether that be for bidding or broadcasting audio files. Finally, latency is not an option for any kind of operation (purchase or delivery): SoundCast requires extremely high availability and reliability to keep things running smoothly.
OPTIMIZED, EASILY MANAGED SCALABILITY
In order to scale up in a cost-effective way, SoundCast has gravitated toward cloud infrastructure from the outset due to its inherent flexibility. At the same time, the company has improved its ability to manage peaks in traffic with software architecture based on microservices and independent containers, both of which can be duplicated as required, depending on the number of requests to be managed at that given moment. « This made Kubernetes the obvious choice to roll out and manage our containers—especially since a load balancer upstream of the Docker engine allows performance to be optimized based on load, and ensures high availability for our services, » explains Christophe Peillet, CTO of SoundCast. « Google developed Kubernetes. This made Google Cloud a logical and optimal choice for cloud infrastructure. »
As well as Kubernetes, SoundCast also uses Cloud Storage to store audio files. Recently, the company has also started to progressively develop new services and to use BigQuery‘s analytical capabilities on the company’s tracking data.
Four years on, Patrice Sguerzi is pleased with these technology choices that have contributed to the company’s growth: « We’ve doubled our revenue every year since the company started. This September, we handled a billion invocations in one month, and the platform was already connected to 10,000 podcasts, 200 live and web radios, and 120 games (audio in gaming) at this point. Things are looking very positive. »
Given our choice of architecture, Kubernetes was the obvious choice. This made Google Cloud a logical and optimal choice to support our development, and manage fluctuations in our network traffic with large peak loads. And to be able to complete this large-scale project, our first thought was a Google Cloud Partner: OP-Rate
ANTHOS MCI FACILITATES INTERNATIONAL EXPANSION
In 2023, SoundCast plans to continue expanding and double its revenue once again. To reach these goals, the company has decided to expand its operations on a global scale. Primarily targeting the US and Asia, SoundCast has duplicated its existing cluster in these two regions so requests can be processed locally to optimize response times.
To accomplish this, the company has been supported by OP-Rate, a Google Cloud partner with significant expertise in the mass processing of feeds. Functioning as an extension of the existing team, OP-Rate works seamlessly with the IT department, fostering a dynamic and agile relationship between the SoundCast and OP-Rate teams. This collaboration has brought together routine tracking and technical optimization so deadlines can be met. « At every stage, OP-Rate’s responses have been exceptionally precise and their commitment has been excellent. We were impressed by their ability to meet tight deadlines without sacrificing quality. The result speaks for itself. The roadmap has changed very little—OP-Rate has been reliable and reassuring at every stage, without forgetting how available the executive and operational teams have been. OP-Rate delivers on every level. We’ve really appreciated the daily communication and support, » explains Patrice Sguerzi.
Today, we're looking to harness Google Cloud's sustainable development strategy and tools in the services we offer, as more and more of our customers are asking us about this topic.
— Patrice Sguerzi, CEO of SoundCast
In order to redirect traffic to the cluster closest to the user, OP-Rate relies on Anthos MCI (Multi Cluster Ingress), a global load-balancing service that automatically selects the closest cluster based on the origin of the request to optimize the response time of SoundCast’s service. What’s more, if the European cluster is down, for example, Anthos MCI can automatically redirect traffic to another cluster in the US or Asia, which means SoundCast can provide greater reliability.
The startup’s next step is to take advantage of Google Cloud’s carbon-neutral status and the many tools associated with Green IT to develop a range of services that contribute to sustainable development. « It’s something we can promote to our customers, particularly now that more and more of them ask us about it, » concludes Patrice Sguerzi.
SoundCast creation is the natural extension of professional backgrounds, personal interests for the media and « sound » in all its forms. Between 10 and 15 years each of experiences in the digital advertising world, it represents qualitative editorial environments, innovative ad solutions, and offers relevant experiences to mains brands.
- Google Cloud
- Kubernetes Engine
- Cloud Load balancing
- Cloud MemoryStore
- Artifact registry
Découvrez nos réussites
de transformation numérique
Modernisez ses applications
Devenez Cloud Native et passez à l'échelle avec Google Kubernetes Engine.
Moderniser son infrastructure
Sécurisez et fiabilisez votre plateforme avec Google Compute Engine et Kubernetes Engine
Moderniser sa plateforme Data
Moderniser votre Data Lake et votre Data Warehouse avec BiqQuery, Looker, Dataflow & Looker