Video Streaming Platform

About the Client

Pluto TV started as an early player in ad-supported video-on-demand. It initially served live channels and an on-screen programming guide and mimicked the traditional TV viewing experience. However, the company achieved a real breakthrough when it added on-demand videos to its content offering in 2017. Thanks to that, users could browse and stream movies, TV shows, and other content anytime.

In January 2019, Pluto TV reached another crucial milestone. The company was acquired by ViacomCBS, a leading global media company. It partners with the top TV networks, digital media creators, сontent providers, and brands to distribute video content directly to viewers over the Internet. Pluto TV’s free, ad-supported offering is available in 25+ countries, with millions of viewers tuning in each month to watch premium news, TV shows, movies, sports, lifestyle, and trending digital series. 

The client generates most of their revenue through video advertising by operating a high-load system, which receives millions of requests per minute. The solution is connected to ad exchanges, networks, data centers, and other services to manage a complete ad sales cycle.

Goal

The client wanted to address future needs and let the system evolve with time. Planning the expansion of their services, Pluto TV wanted to optimize the system’s performance through technology.

The system had a monolithic architecture that contained large volumes of legacy code. It made it complicated to introduce new functionality, improve scalability, and optimize the required performance. A solution-to-be was expected to cope with a continuously growing amount of viewers, track ad impressions, and analyze the efficiency and relevance of advertising campaigns. 

Without deep in-house AdTech expertise, the client searched for a highly qualified and experienced team to implement the project. Geomotiv became an ideal candidate for this project thanks to extensive experience in Linear and OTT TV and adjacent verticals like Adtech. Here’s what it looks like in more detail.

Our team was engaged in developing a custom linear TV media selling product that helps track linear TV ad campaigns in real-time. Another notable project was dedicated to creating a programmatic video distribution platform for a media and advertising industry leader.

We believe our experience with developing custom platforms for OTT and Linear TV enabled us to become one of the most proficient tech companies in these domains. It culminated in creating a full-stack programmatic AdTech platform for streaming media services, publishers, vMVPDs, media agencies, OTT and AdTech technology vendors, and startups.

Work Description

Duration:
2017- ongoing

Team:
Upon the project’s start in mid-2017, we provided a dedicated team consisting of 3 specialists. They quickly gained an understanding of our client’s requirements, technology needs, and prospective solution architecture. The client was impressed by the initial results and they wanted to hire a few specialists full-time. A new team was gathered to develop their advertising technology. As a result, the team strength on our side rose from 3 to 10 people within the first year of our cooperation.

In the next three years, the team size grew exponentially. We currently operate a couple of mixed sub-teams, including our developers and QA specialists and our client’s in-house employees. In addition, there is a team that is fully managed on our side as well. 

As the project expands every year, the client is attracting new hires and forming new departments to operate their OTT service and its parts. Following success in the market, Geomotiv scales up its development capacities according to the client’s needs.

Approach:
Together with Pluto TV specialists, we decided to opt for the Agile methodology with two-week development sprints. We chose such an approach because it allowed us to cope with more tasks and make the work efficient. 

We foster close collaboration between the Product Owner and the development team. The regular activities include sprint planings, backlog grooming, retrospectives, and daily stand-ups. There are also traditional Scrum of Scrums meetings that aim to foster interaction between teams and ensure alignment.

Services provided:
Since our client aimed to optimize expenses and introduce new changes to the system to pave the way for future expansion, the team was tasked with its consecutive upgrade. Most of the legacy code in the system was written in JavaScript from 2013 to 2015. To stay in tune with the commonly used tech stack, we broke large chunks of the monolith into microservices written in TypeScript. Subsequently, the team gave up on the monolith and started creating and deploying microservices in the cloud. This enabled the team to detect more bugs at the development stage before moving to production. 

We used Go to write microservices with increasing performance requirements and a small amount of business logic. It enabled us to achieve 10x performance gains in some cases. In other words, we were able to reduce server costs by ten times, which was the need of the hour. As for infrastructural changes, we are on the way to utilizing cloud-agnostic architecture and doing away with Amazon-tied services. Instead, we transfer parts of the system to the Kubernetes stack and employ service-mesh architecture step-by-step.

Our team also implements Header bidding instead of the Waterfall approach using the above tech stack. Finally, one of the most critical components we built for Pluto is a robust ad optimization service that dramatically increases fill and revenue and, at the same time, decreases infrastructure costs.

We implement an end-to-end approach to testing and quality assurance. Our team initiates testing activities at the requirements analysis stage conducted by a BA or a Product owner. At this stage, we:

• analyze requirements;
• calculate potential risks;
• provide initial estimates.

At the stage of producing a working prototype, our specialists:

• draft test cases using a Test case management tool (we use TestRail);
• carry out manual testing according to test cases;
• write automated test cases if there is a real opportunity to automate this functionality (depending on the service, the real-life level of auto-test coverage varies from 20 to 80% of the total functionality);
• update/delete (if necessary) existing automated test cases;
• update (if required) load tests.

Before release we:

• run automated test cases;
• run load testing;
• check manually and selectively the critical functionality of our service.

Work Stages

1. New architecture design
2. Backend development and other tasks
3. Code review
4. Deployment and DevOps
5. End-to-end QA automation
6. SCRUM

Results

By leveraging Geomotiv’s dedicated team model, the client could access deep AdTech expertise and acquire a long-term strategic partner. In addition, the high technical excellence of our developers enabled the client to significantly reduce the costs of the introduction of the new functionality.

Together, we reached crucial milestones that enabled our client to stand out on the market. Our client managed to expand the geography of its services to an impressive list of countries. There has been exponential growth since 2017 from these perspectives as well:

• 5x dedicated dev team growth;
• 3x revenue growth;
• 4x ad impressions growth.