Social Media Ad optimizer
Client
An American start-up looking for a solution based on large data sets that allows for quick processing of advertising offers
Expectations
The client expected a solution enabling the automation of the implementation of advertising campaigns on the Internet and physical platforms such as smart TVs, advertisements in taxis or at cash registers, along with the precise definition of target groups. The solution was to enable offering in each customer segment (from SME to corporations) and granulation of target groups, as well as full traceability of the spent funds.
Project
client application (React, Redux)
backend of the client application (Scala, Akka-http, Akka-streams)
analytical module (Spark, Hadoop, Hive, Athena)
decentralized "Pixel-server" (Scala, Akka-http, Redis) which is its own tool for monitoring ad display in real time
Due to different needs and optimization of efficiency and costs, data can be stored in various ways: databases, FTP servers, HDFS, S3. The dynamic and multi-directional development of the project required the use of good design practices and tools which support the quick and safe introduction of changes, such as: microservice architecture, code review, unit tests, automated tests eg. Jenkins, Liquibase, Docker. Business processes are fully automated as well as managed and monitored in one central, dedicated tool. The entire platform, on the other hand, runs in the AWS cloud, using machines in several geographic regions. This allows for easy and cost-effective management of computing resources and their continuous monitoring.
Results
Thanks to the implementation, the client has a cloud-based solution with all the tools they need to conduct multi-platform advertising campaigns himself, while precisely measuring their effectiveness. This means they can rest assured that their ads are finding the right people and that their budget will not be wasted. At the same time the solution is available as B2B ‘white-label’ product.
Info
Sector
Retail
Service
Produkty dedykowane,
Cloud
Technologies
Scala, React
Redut, Spark, Hadoop
