Unified Content Intelligence to Inform Executive Decisions
SUMMARY
Our client, a massive media conglomerate, owns a vast array of content across their diverse brands and needed a way to understand their catalog to make informed business decisions.
Their ML team was on a tight timeline to demonstrate compelling results toward building a unified content intelligence layer that could work across the catalog.
Hop was engaged to help identify and evaluate ML/AI approaches aligned with their vision and criteria.
Hop worked closely with the content teams to incorporate their processes and insights, ultimately demonstrating what’s possible and answering the question of whether or not further investment would provide value.
THE COMPANY
Our client is one of the largest media conglomerates in the world, producing and distributing a wide range of entertainment, sports, and news content. Hop Labs originally established a partnership with this organization through a strategic engagement that led to further collaboration.
THE CHALLENGE
As one of the largest media conglomerates today, our client encompasses a diverse set of brands, cycles, production processes, and supply chains. To remain nimble in this fast-moving industry, they needed a unified content intelligence layer that worked across their content catalog to inform monetization approaches. Our client’s business teams had already identified specific dimensions for their content, and asked Hop to help identify ML/AI approaches toward understanding their asset catalog across those dimensions.
A challenge tangential to the core ask was a tight timeline to demonstrate compelling results. As with any successful large enterprise, communicating with the necessary parties and accessing data can be complex, and our client counterparts helped to pave the way for our team to deliver something useful and effective within a matter of months.
“This project has been hugely successful, and our leadership’s response has been extremely positive. Under suboptimal circumstances, your team clearly exceeded expectations.”
THE APPROACH
A variety of approaches exist in the literature with regard to applying ML for content intelligence; however, not all of these are aligned with our client’s vision. They had established their own business criteria for “high-quality” solutions, and for any AI product in their stack. With this in mind, our approach was to get started quickly with an initial dataset to establish a rigorous baseline, doing the simplest thing to explore and quickly understand the space. This enabled evaluation of techniques from the literature, prioritized as per our client’s criteria. The content teams then added their nuanced insights, leading us to reprioritize the techniques to pursue further. What evolved was a two-pronged approach – our team continued to move forward with scientific findings on the initial limited dataset, while simultaneously validating prior findings on larger and more robust datasets.
“My number one reason for recommending Hop’s services is the ability to get through the noise and get down to the R&D phase. There’s often quite a bit of fluff in working with consultants, but Hop quickly gets to cranking out code, research and metrics.”
Although we were able to work with our client to assemble various datasets from their existing systems, these datasets were not labeled at the level of granularity necessary for a well-posed ML problem. This required our scientific team to creatively reformulate the relevant ML problems in order to leverage the human insights baked into the existing organic labels. Looking forward, we’ve coordinated with our client’s executive team to prioritize deeper labeling efforts to leverage human feedback and insights in an ongoing, strategic way – the foundation for a subsequent project to align operational workflows.
THE RESULTS
While we’ve only begun the work toward applying ML to improve our client’s content intelligence, this short engagement quickly demonstrated results on the lower bound of what’s possible, answering the question of whether or not there could be real value here for their organization. With a clearer vision of the potential for a return on investment and competitive advantage in the industry, the executive team is continuing to move forward on this path. We look forward to collaborating with them on the next phase of work.
Need a better understanding of your data to inform business decisions? Contact us to learn how Hop can help.