Data in the Driver’s Seat

3 things video publishers can learn from Group Nine Media at JW Insights

Home to content brands like Thrillist and NowThis, Group Nine Media has established itself as one of the world’s most successful video publishers. Each month, the digital media holding company receives about five billion views across its portfolio of premium brands, which also includes The Dodo and Seeker.

At  JW Insights, Variety co-Editor-in-Chief Andrew Wallenstein joined Group Nine Media CEO Ben Lerer and Chief Insights Officer Ashish Patel for a discussion of the role data has played in the firm’s growth. Here are three things we learned.

 

1) Publishers can learn as much from failure as they can from success.

While many publishers typically look to replicate their most successful content, Group Nine Media spends more time analyzing why unsuccessful content didn’t work.

 

By looking at dropoffs during its most poorly viewed videos one month, Group Nine discovered that mobile viewers were being turned off by the way certain text was formatted on the page. When it tested videos with the text reformatted, it quickly saw better results.

 

“We simply believe that virality is too variable to be a replicable strategy, so a lot of the focus of our learnings is more so on the bottom of our [content] stack, on what didn’t work as opposed to what worked.” Patel said.

 

2) Group Nine is working to generate insights for videos before they’re even published.

Through its Group Nine Insights Analysts program, Group Nine Media is building technology designed to automate the data-based recommendations it gives content creators.

 

Already, the firm’s technology uses computer vision to automatically create metadata that shows what’s happening onscreen at any point in one of its videos. By comparing the creative elements of a video against the performance of past videos with similar elements, Group Nine is now able to predict its retention curve — a graph that shows where viewers stop watching a given video — within a 5-10 percent margin.

 

Ultimately, it hopes to use its technology to build an automated process that gives content creators recommendations for optimizing a video before it’s published. “We think that this is the way that we can start to scale our process and allow ourselves to continuously pace against the market,” Patel said.

3) In the era of big data, human creativity is still essential.

As sophisticated as its data operation is, Group Nine still relies a great deal on the instincts of its creators.

 

“We want to use the data to position the stories that content creators want to tell better, as opposed to telling them what to say,” Patel said.

 

Lerer added that publishers who allow data to hold too much sway over their creatives can very quickly find themselves turning into a “content farm,” noting that many publishers who tailored their content to the Facebook algorithm lost reach when the algorithm changed.

 

“They’re doing the data part, but the data part in absence of real serious context, real editorial credibility, and real great storytelling is just — there’s nothing there,” Lerer said. “Generally speaking these [algorithm changes] end up being positive moves for us, not negative ones, because they separate the wheat from the chaff and they allow the mixture that we have to win out ultimately.”

 

 

To watch the full session:

 

 
For more from the JW Insights blog series, click here.

 
Want to start making the most of your video data intelligence? JW Player’s video analytics puts real-time information at your fingertips, allowing you to optimize your advertising and content.  Schedule time to speak with a video expert.

 

Contact Us

The post Data in the Driver’s Seat appeared first on JW Player.