Syracuse University’s Local NExT Lab argues that dashboards cannot show causes, and its first experiments with Mississippi Today and New Orleans Public Radio found surprising results about story context and Instagram
Dashboards Are Not Enough
Newsrooms today are awash in data. Analytics dashboards track page views, time on page, scroll depth, social shares and subscription conversions in real time. But data about what happened is not the same as knowledge about why it happened. That is the central argument of a new report from the Local NExT Lab at Syracuse University, described by Nieman Lab deputy editor Sarah Scire in a piece headlined a new report argues newsrooms need to do more experiments, the kind with randomization and results, published on October 15.
The report, the lab’s Year 1 Impact Report, puts it succinctly. Dashboards and engagement analytics are useful, but they cannot show causes. Randomization can.
Experiment One: Story Context
One of the lab’s first experiments was conducted with Mississippi Today, the nonprofit newsroom whose investigative reporting has won a Pulitzer Prize. The question was whether the stories surrounding a major investigation changed how audiences received it. A national survey of 1,768 adults saw the same investigation presented among different mixes of other stories.
The results were striking. The investigation was rated more accurate, more important and more relevant when it was surrounded by community-focused stories than when it was surrounded by other investigations. A mix featuring stories about legislation produced inconsistent reactions. The implication is that how a newsroom packages and presents its work, not just the work itself, shapes how audiences judge its credibility.
Experiment Two: Faces Versus Brands
A second experiment, with New Orleans Public Radio, tested approaches on Instagram. Reels featuring reporters’ faces drew the most interest and engagement. But when it came to following an account, profiles that looked like traditional news brands outperformed accounts named after individual reporters. In other words, audiences responded to people in content, but trusted institutions when deciding whom to follow. That nuance would be difficult to discover from engagement dashboards alone, which tend to reward whatever generates the most immediate reaction.
A Low-Cost Partnership Model
The lab’s approach is designed to make experimentation accessible to newsrooms without large research budgets. Newsrooms commit to only two or three conversations. Researchers design, run and analyze the tests. Newsrooms receive causal evidence about their own ideas without spending money or staff time. The first tests were funded by the W.K. Kellogg Foundation, and the American Journalism Project will support further tests with a larger cohort next year. Joshua Darr, the lab’s director and an associate professor at Syracuse, is quoted in Nieman Lab’s coverage.
Why Newsrooms Rarely Experiment
Randomized experiments are standard practice in medicine, economics and technology. Large technology companies run thousands of A and B tests to refine products. Yet most newsrooms make decisions about story presentation, headlines, social strategy and product design based on intuition, precedent or correlational data. There are good reasons for that: experiments require expertise, time and statistical understanding that small newsrooms often lack. But the cost of guessing wrong, in wasted effort or lost trust, can be high.
There is also a cultural factor. Journalism has a long tradition of editorial judgment, and some editors are wary of letting experiments override instinct. The Local NExT Lab’s argument is not that data should replace judgment, but that judgment should be informed by evidence about what actually works.
Data Done Well, Commercially
The same principle, using data carefully rather than simply collecting it, is visible on the commercial side of the industry. Press Gazette reported on October 15 that the Guardian boosted its programmatic advertising revenue by 21 percent in a year through a project called Curation at Scale, which simplified hundreds of pricing rules, removed intermediaries and offered advertisers more detailed audience signals, such as a visit to a sports page, using privacy-safe technology that did not hand over readers’ personal identities. The project’s architect warned that it was not a template to copy, and advised publishers to understand their own audiences and supply chains first. The lesson echoes the Syracuse report: data is most useful when it answers a specific question.
Beyond The Buzzwords
The media industry is not short of jargon about data-driven strategy, audience-first thinking and optimization. The London Prat captured the absurdity of some of that language in a story about how a PR company became the first agency to generate more buzzwords than revenue. Randomized experiments are the antidote to buzzwords. They replace claims about what works with evidence about what works, and they can produce answers that surprise even experienced editors.
The findings also have practical implications for editors right now. The Mississippi Today experiment suggests that pairing hard-hitting investigations with community-focused stories may increase readers’ perceptions of their credibility. The New Orleans experiment suggests that newsrooms might feature reporters prominently in video while keeping institutional brand accounts as the main home for followers. Neither finding is a universal rule, but both are testable ideas other newsrooms could explore.
What Comes Next
The lab’s expansion next year, with support from the American Journalism Project, will test whether its model can scale. If it does, local and nonprofit newsrooms could gain access to rigorous evidence about how to present their journalism, build audiences and earn trust, without the research budgets of large media companies. In an industry facing falling traffic, shrinking staffs and declining public trust, knowing what actually works, rather than guessing, may prove to be one of the most valuable tools a newsroom can have.