Google Play and YouTube Movies & TV Content Platform (2018)

I joined Google in December 2017. I had been recruited to join the Google Play 1st Party Media & Entertainment Operations team as an Operations Manager. The team was doing a lot of cool things in overhauling anything from partner programs, content aggregation and standardization, connected services and integrations, to creating a next generation platform. A few months into this new gig the technical program manager for Play Engineering left for another role, and I was asked to step in as a “loan” to the product and engineering teams.

It was tricky, and intimidating, not because I wasn’t confident that I could do the role, but because I was an outsider that was new to Google. I was stepping into a role filled by a very likeable guy and fellow engineer. I was a woman (when 85% of the org was men). I had no engineering or computer science degree (TPMs are required to have one of these and be able to code). I was also coming into a project that had a failed history (a previous iteration that was two years in the making had crashed and burned). And lastly, I had little to no authority (but I did have support) to completely overhaul how the engineering, product and operations teams developed and executed this project, and to actually ship the project.

We had big milestones to hit with some seriously lofty goals and a relatively junior (new hires to Google and new to media and entertainment) engineers, product managers, and operations program managers.

What was the project? We were redesigning the movies & TV content ingestion pipeline and overhauling the legacy system infrastructure for Google Play and YouTube. The goal was to stand up a platform to enable content at scale, and develop features that made complexity simple for commerce, discovery and playback.

The project included: standing up a brand new component based and industry standardized e2e content ingestion platform, implementing new content aggregation and standardization requirements and policy and API for 3rd parties, the creation of a new transcoding and encoding technology pipeline on GCP (Google Cloud Platform) to support 4K UHD and Dolby Vision, deploying a new partner facing CMS (content management system) Play Movies & TV Partner Portal for publishers, post production houses, quality control, rights management companies and Googlers to manage workflows and assets, deprecating outdated legacy YouTube tools and breaking dependencies across the ecosystem, developing and consolidating vendor operations worldwide, and developing and shipping new capabilities like playable sequences, multi-audio track, and language holdbacks.

All-in-all it was a massive undertaking and would impact ~400 of our global media and entertainment partners across 41 countries, and users of Google Play and YouTube everywhere, including the 1.4B Android users worldwide.

Despite the massive milestones to ship we were making changes in-house simultaneously. We drove the engineering, product, UX and operations teams to adopt Agile. We adopted software project-planning tools that helped the team manage bugs and sprints in order to form realistic expectations about when work might be completed based on the team's ongoing performance. We moved from large phase milestones to feature based. We re-wrote business requirements docs and adopted user stories. We developed and implemented user experience design principles in order to evolve usability of our portal, and consolidated ~23 weekly meetings down to one week cross functional program meeting.

None of these were easy tasks, and we experienced a lot of hurdles along the way. It takes time to make organizational changes (like the above) and to gain the trust of multiple teams that have very different organizational goals. We spent a lot of time doing just that, building trust. We used Google and non-Google methods and tools to get there. We committed to work together to develop solutions, and to make compromises and decisions as “one” team. In the end all of the work that we did made us a very strong virtual team, and it paid off.

We decided to do the tricky implementation and support both the legacy and next generation pipelines simultaneously. We launched the NextGen pipeline in April 2018. By the end of the year we’d migrated 10% of our movie partners to the platform with massive success. It was a tricky play. Our approach was to migrate partners with low volume and revenue numbers and build up migration over time to larger major partners, as more complex features shipped.

We started working the mid- and major- studios with more complex IP and workflows to be our partners. We came to them and said “give us feedback, do user testing, be part of the process”, and they did. Our partners input was invaluable and led to critical changes in the overall vision and execution. By the end of 2019, ~80% of our movie partners had migrated the new platform (with TV migrating in 2020).

The people and teams that contributed to this massive undertaking were outstanding. Everyone rose to the occasion and really committed to developing a scalable component based and industry standardized movies and TV content ingestion platform.