Finding the Right Tool to Accelerate Output
SUMMARY
Our client’s data science team needed help with evaluating options for accelerating their output.
This evaluation required Hop to first gain a deep understanding of current needs and pain points across all stakeholders.
With an informed approach, Hop was able to steer our client away from what would have been a poor investment in an existing tool and assist in building an internal ML platform from scratch.
Our client ended up with a reliable, open-source platform that resulted in an 8-10x acceleration of the output of their data science team.
THE COMPANY
Our client is one of the largest news outlets in the world, a trusted brand with huge impact in the U.S. and beyond. As with any other major media company reaching millions of viewers per day, they are challenged with charting a course into the future of media in an internet-driven world with decreased attention span. Particular to this organization was the question of how to move beyond their reputation as a go-to source on television for sporadic breaking news toward ongoing engagement via various platforms with their audience. The executive team narrowed in on developing and delivering personalized content to achieve this goal.
THE CHALLENGE
When Hop was engaged, our client’s internal data science team had begun the work of developing personalized content, yet were experiencing challenges in getting that work into production. They were in the midst of exploring tools to accelerate the output of their team and needed Hop’s help with evaluating the options available. What tool did they need? Would it be better to build a tool or buy an existing one?
THE APPROACH
Our team’s first step was strategic: we needed to gain a current understanding from all stakeholders – their needs, their pain points, the bottlenecks in existing processes, and so on. From there we had an informed approach toward evaluating the value of existing platforms on the market and the potential of building a tool from scratch. At the time, our client was contemplating a major investment in a particular tool that Hop was able to critically evaluate against their needs, ultimately determining that it would not be a good fit.
With our findings that no existing tool was an ideal fit for our client’s particular needs, the best path forward was to build an internal platform. Hop assisted with engineering and implementing an ML platform from open-source components.
“As an informal estimate, our data science team believes they were able to test twice as many models in Q1 2021 as they did in all of 2020, with simple experiments that would have taken a week now taking half a day.”
THE RESULTS
Having avoided a poor investment, our client ended up with a reliable, open-source platform that they were free to extend and avoided vendor lock-in. Utilizing the new tool resulted in an 8-10x acceleration of the output of their data science team. More details about the creation of this platform can be found in this Medium article.
Need an informed perspective on a key strategic decision? Contact us to learn how Hop can help.