Elasticiti helps popular brand within a network solve big data questions on a short timeframe


 

Challenge:

With big data infrastructure still 12 months away, this network needed help analyzing big data fast

Solutions:

  • Data munging in R

  • Audience-overlap matrix in Tableau

Length of Engagement: 2 months

 

Background:

This famous international brand wanted to track individual user behavior -- across shows and over the course of a 12-month period -- so the marketing team could understand the strengths between different audiences and different properties. Of particular interest were instances where different shows that had strong commonalities in audience.

By understanding which shows had common audiences, the brand could drive efficiency in its tune-in and marketing initiatives, as well inform its content-creation and distribution strategies.

The brand was one of many in a large network which had grown both organically and via acquisitions. As a result, individual brands are rather siloed and access to corporate resources was often limited.

 

Challenge

Thanks to its Adobe Omniture system, the analytics team has access to detailed user data. The challenge was that each month contained three gigabytes or more of user data and the quarter contained almost eight. The team wanted a full year’s worth of analysis to get the best picture of audience overlap.

Clearly the team needed a Big Data infrastructure – and its parent company was busy building one. Unfortunately for the team, its brand wasn’t scheduled for onboarding for another 12 months.

The team tried built an audience-overlap matrix in Microsoft Excel, doing the aggregations in Omniture but it presented several challenges: it was time consuming, offered limited ability to slice and dice the data, and it didn’t help them look at their data over time. As a result, they were locked into a high-level summarized view of their data.

 

Solutions

The network engaged Elasticiti, which set three goals for the project:

  1. Create a proof-of-concept using two full months of user data

  2. Design an audience-overlap matrix that could be viewed and manipulated on laptops, enabling a wide swath of people within the network to use them

  3. Implement a transition strategy that would allow the in-house analysts to complete the remaining ten months of data themselves

Elasticiti used R to do the data munging and overlap analysis, a program capable of ingesting large datasets and performing the calculations needed to understand the audience overlap.

The data was then fed into Tableau, a program that offers the rich visualization capabilities to display a robust audience-overlap matrix.

“It was important to us that the matrix run on a laptop so it could be used widely within the organization,” Ben Reid, CEO of Elasticiti explained. “A lot of research groups work closely with the site and editorial teams, and this was certainly the case with this client. The overlap analysis helped the site team in design. It also helped the editorial get insight back to the content producers. That was pretty central to the project.”

Finally, Elasticiti documented every step of the process, and trained the client’s internal analysts on how to import the data into R, run the calculations, and create the matrixes in Tableau.

 

Results

  • Deep understanding of audience overlap. With a deep understanding of its shows, the network is now able to discern which signals are meaningful and should be pursued. This has allowed the marketing team to drive significant efficiency in their efforts.

  • Immediate solution in place.The network now has a complete solution to perform the analysis as they wait for the parent company to build the next-generation of infrastructure.

  • Shared insights among internal teams. Because the Tableau matrixes are easy to use, the data and insights they contain are now freely available to the editorial team and content partners.

  • Two-month engagement and hand-off. Perhaps most important of all, by taking over the project, the brand’s team got smarter and learned new data skills that make them more self-sufficient.

Questions about how we can help you?