Demand-Side Platform (DSP) Development for Outfit7 

Introduction 

Outfit7, one of the most well-known companies in the mobile-gaming industry, asked us to build a custom demand-side platform (DSP) designed specifically for integrating advertisements into its gaming applications. 

At the time, Outfit7 was using an in-house ad serving system, but wanted to replace it with a custom-built DSP.  

Outfit7 decided to transition towards a cross-promotional model, essentially aiming to promote their own games inside its existing apps. This plan was formulated to introduce its new games to its existing users.  

Avenga assisted Outfit7 in honing the focus of the MVP phase. We were involved in helping Outfit7 build a platform designed to enable the promotion of its games across its catalog of games. 

Challenge 

Outfit7 was relying on an in-house ad serving system that no longer met their needs. They lacked control over the existing platform, which limited their ability to adapt it for promoting new games effectively. The company wanted to shift toward a cross-promotion model, targeting their own user base within their catalog of mobile apps. They needed a solution that offered precise ad decisioning, a bidding-based approach, and full ownership over the technology, with flexibility to evolve as their business and partnerships grew. 

Solution 

Process 

User Interface and Campaign Management

We designed and developed an intuitive user interface to simplify campaign setup and management.

Alongside this, we created a comprehensive campaign management system that supports multiple ad formats to meet varied promotion needs.

API and Bidder Development 

Our team developed a dedicated API to enable seamless communication between the DSP and the bidding components.

We also built a high-performance bidder responsible for placing bids and purchasing ad impressions in real time.

Tracking and Metrics Exporting

We implemented a tracking system that captures key events like impressions, conversions, and in-game actions to measure campaign effectiveness.

Additionally, a metrics exporter was developed to provide detailed performance reports and analytics.

Creative Optimization with Reinforcement Learning 

To maximize ad effectiveness, we developed a creative optimization mechanism using a reinforcement-learning algorithm called the Softmax Multi-Armed Bandit.

This algorithm analyzes impressions and conversions to select the highest-performing ad creatives automatically, improving campaign results over time.

Results 

• We delivered a working MVP to Outfit7 that included all the required features and functionalities. 
• Outfit7 is testing the DSP internally. 

With the custom DSP now in testing, Outfit7 has gained full control over its game promotion with precise campaign management and optimized ad performance. This tailored solution replaces their legacy system, delivering flexibility, transparency, and advanced tools to support future growth. The project shows how custom-built AdTech platforms can empower businesses to own their marketing and drive results.