Nieman Foundation at Harvard
HOME
          
LATEST STORY
From shrimp Jesus to fake self-portraits, AI-generated images have become the latest form of social media spam
ABOUT                    SUBSCRIBE
Nov. 2, 2020, 8:30 a.m.
Reporting & Production

With Talk2020, The Wall Street Journal turns an internal reporting tool into a reusable news product

During a fraught election season, a transcript database is reimagined as a news product for readers reaching for the right quote “to make a point.”

When I was growing up, every holiday meal was interrupted, at least once, by someone leaving the table to gather supporting materials — a dictionary, the atlas, a photo album — to make a point. Plates were pushed aside and an index finger would thunder down in triumph. “See?!”

We tend to reach for our smartphones or another Google-enabled device now, of course, but that impulse to marshal evidence as we discuss and debate with our loved ones lives on. And it’s one of the reasons The Wall Street Journal designed and debuted Talk2020, an experimental tool that allows readers to search a database of thousands of transcripts to see what the presidential candidates have said about an issue.

Through extensive focus groups conducted last summer, the Journal found that readers wanted to be able to quickly locate quotes and facts about a candidate’s record — and not just for their own edification.

“One of the things we know about our audience is they’re frequently working to get support for points that they would like to make to folks in their lives,” senior program manager Tania Feliz said. “This tool allows them to find specific information and to share it.”

Users can filter by issue, date, candidate, or keyword. A coworker who cares deeply about Palestine? A father-in-law who heard that Biden was going to ban fracking? A friend who wants to know if Trump really said he wanted to slow down coronavirus testing? (“See?!”)

The transcripts are pulled from campaign speeches, media appearances, debates, and more. (The Journal added Kamala Harris and Mike Pence to the list after their debate made it clear just how much interest there was in the candidates vying to be one heartbeat away from the presidency.)

Becky Bowers, a D.C.-based strategy editor for the Journal, explained that Talk2020 began as an internal tool used by reporters and editors who work in the Washington bureau. With Dow Jones’s Factiva as a backbone, Talk2020 began as a searchable database of President Trump’s speeches custom-built for D.C. correspondents — one that they’ve been using to frame and inform their own journalism since 2019.

Feliz, along with product director Tyler Chance, helped transform the internal tool into a product useful for readers. “We had the data. What we didn’t have was the product,” Bowers said. “How can we take what we have and reframe it in a way that meets audience needs, looking at their habits and need to break through the noise?”

Chance emphasized that Talk2020 users seemed to appreciate getting the “raw materials” and “primary documents” from a trusted news source. He said that, in particular, readers liked to “catch up” on what was said during debates — even if they had tuned out for the actual broadcast.

The Journal identified that “catch up” need before. In March, the Journal’s chief news strategist and chief product and technology officer Louise Story explained it was behind one of the tools designed for political coverage and elections that got repurposed for pandemic coverage. (Bloomberg Media has since implemented a similar feature for mobile web users — called “All Caught Up” — that allows readers to swipe through article summaries to quickly catch up on major news stories. Bloomberg’s version uses artificial intelligence technology, rather than the Journal’s editors, to select and summarize the news.)

Launched as a prominent homepage feature, Story said the early response to the Journal’s “catch up module” indicated that the users visiting wsj.com directly were not the “main audience” looking for new summaries; instead, they were the ones more likely to be reading all — or, at least, many — full articles.

No matter, said Story. (A wonderful aptronym for a journalist, no?) She says the Wall Street Journal is confident in their user research and plans to retool the way they deliver the catch-up module to find the right segment of their audience.

Sarah Scire is deputy editor of Nieman Lab. You can reach her via email (sarah_scire@harvard.edu), Twitter DM (@SarahScire), or Signal (+1 617-299-1821).
POSTED     Nov. 2, 2020, 8:30 a.m.
SEE MORE ON Reporting & Production
Show tags
 
Join the 60,000 who get the freshest future-of-journalism news in our daily email.
From shrimp Jesus to fake self-portraits, AI-generated images have become the latest form of social media spam
Within days of visiting the pages — and without commenting on, liking, or following any of the material — Facebook’s algorithm recommended reams of other AI-generated content.
What journalists and independent creators can learn from each other
“The question is not about the topics but how you approach the topics.”
Deepfake detection improves when using algorithms that are more aware of demographic diversity
“Our research addresses deepfake detection algorithms’ fairness, rather than just attempting to balance the data. It offers a new approach to algorithm design that considers demographic fairness as a core aspect.”