Demand-side platform in numbers


end users


requests per second


GB of DSP data daily

Challenge: Core platform locked in legacy architecture and outdated UI

Europe’s leading demand-side platform that enables digital advertisers and marketing agencies to effectively manage campaigns across multiple ad exchanges and in real time.

DSP provider: “Maintaining a leading position is hard as it is, with an obsolete tech stack it is impossible.”

“We are leaving money on the table because of our core platform’s architectural limitations that make it a nightmare to grow our customer base while also trying to deliver stable performance. And the complex business logic often stands in the way of adding new features efficiently to stay ahead of the competition.”

Rearchitectured solution: One application, one database

With the old architecture, registering and managing each client required launching a separate application instance, which resulted in enormous resource consumption and poor scalability.

To solve this issue, the project team rebuilt the platform — all client entities are now stored in a single database and managed via a single admin panel, which saves server resources and optimizes their usage. The broad scope of the admin panel functions includes:

  • Editing clients’ details (name, logo, geotargeting, and more)
  • Adding user accounts and defining their rights
  • Assigning inventory types
  • Managing third-party data providers for precise audience targeting

Big data expertise with the focus on costs optimization

Before Oxagile’s involvement, the solution relied on Google BigQuery to handle the growing amount of DSP data, which started to cost the client an arm and a leg.

“When you have to process hundreds of gigabytes of ad campaigns statistics, BigQuery storage and query costs can take a toll. To help the client keep costs at bay, we migrated workloads from BigQuery to ClickHouse.”

— Ilya, Ruby Engineer

In addition, Oxagile’s team worked closely with data science experts to enable forecasting capabilities. This will allow digital marketers to predict how many times their ads will be shown and fine-tune their campaigns accordingly.

Catering to the needs of marketing agencies with a streamlined UI

Recognizing the importance of a user-friendly frontend part of the platform, the client tasked Oxagile to build a modern, more intuitive UI that would simplify complex workflows and save time for digital marketers.

Upgrading the user interface required updated business logic and sharp UX decisions. Keeping end users in mind, Oxagile’s team delivered a new UI that streamlined day-to-day operations like ad campaign creation. Now, instead of filling in 50 input fields a marketer can set up primary and secondary ad campaign goals and the system automatically pre-populates the necessary fields, making the entire process more efficient.

Other user-facing features include:

  • Sophisticated behavior-driven ad delivery
  • Deeply customizable ad campaigns, including media plans, start and end dates, inventory sources, budgets and daily spend, bids, and more
  • Multimedia creatives management
  • Batch editing of campaigns and creatives
  • Transparent tracking of changes made to client entities
  • Dynamic RTB seat management

Stepping up the DSP game: Advanced reporting

Detailed reports with new metrics, filters and totals calculations empower advertising agencies to drill down on campaign performance and better understand what drives conversions.

Reporting metrics include:

  • Impressions, clicks, and CTR
  • Remaining media budget
  • Gross price and campaign cost
  • Anomaly clicks
  • Advertisers, data providers, publishers
  • Inventory research, publishers
  • Conversion pixels
What’s more, custom integration with Google Data Studio allows digital advertisers to create powerful, easy-to-understand data visualizations with just a few clicks.
Professional Services
Delivery Model
SLA-driven task backlog delivery (TM-based)
Effort and Duration
9 months, 19 man-months
Ruby, Ruby on Rails, JavaScript, Angular, Apache, Passenger, Percona Server, Linux, AWS, CDN, BigQuery, ClickHouse, Google Memorystore