Did we act fast enough?
It will be one of the most searing and studied questions in the wake of the Covid-19 pandemic. Already, there’s plenty of blame to go around. When the dust settles, few will be absolved of responsibility. Expect nothing less from a tragically pivotal moment in human history.
With this report, we present an additional point of consideration: How did the news media perform when we needed it most?
Given our agency’s daily collaboration with journalists and our fluency in information design and data journalism, we feel uniquely suited to analyze and interpret Covid-19 coverage in the early months of 2020.
To do this, we’ve used the extraordinary Internet Archive to study snapshots of coronavirus-related news and opinion, disseminated over the first 10 weeks of this year, across 18 of the nation’s most prominent online publications. With this data, we constructed timelines of how quickly or slowly these publications pursued the subject, and how each publication’s coverage relates to its peers.
What was counted: For each publication, we counted coronavirus-related headlines, published on the homepage, across a 10-day sample spanning 10 weeks. For details about what qualified, see our methodology at the bottom of this report.
When it was counted: Using the Internet Archive—which each day archives live snapshots of hundreds of millions of web pages—we were able to retroactively analyze coverage as it happened in real-time, on each site.
We began with the first week of January, when coverage in the US was nearly nonexistent, and followed through March 9th, just as the story broke open and consumed the entirety of the nation’s attention. To create a fair sample size across the publications, we measured coverage at a consistent point each week: Mondays between 11 am and 2 pm CST.
Below is our first representation of the data: Total instances of coronavirus-related coverage, above the fold, over 10 weeks.
A headline that’s “above the fold” means it’s visible on a standard-sized computer without the user needing to scroll down. Put more plainly, running a story above the fold means editors deemed it more important than most other stories.
If you want an early scoop, follow the money.
No sector of the mainstream media was quicker and more vigilant in tracking the emergence of Covid-19 than finance-focused publications like Bloomberg, Reuters and The Wall Street Journal. They ran stories early, ran them often, and they were most consistent in getting coverage to the top of the page, in front of the most readers.
By the end of January 2020, the three aforementioned publications accounted for 46 percent of all headlines counted.
With this modest, mid-page story, it all began (at least within the bounds of our analysis). On January 6th, nestled next to stories about Bed Bath & Beyond’s real estate and a tech-savvy move by the indomitable Little Caesers Pizza, The Wall Street Journal introduces “China’s Mystery Virus.”
Given what we know now about how the crisis accelerated and amplified, it’s clear that passive, “below the fold” mentions of some health scare abroad were never going to be enough.
What set a few publications apart as this story emerged was the editorial instinct to know where this might go, and to honor that potential by running it as a lead story. The chart below documents instances of coronavirus-related leads over the first 10 weeks of this year.
In this case, we define a lead as being the first and most prominently displayed story on a publication’s homepage.
Bloomberg and Reuters led the charge in late January, while The New York Times and ABC News had the strongest finishes, each with four consecutive weeks of lead stories heading into mid-March.
In the chart below, we assess the data from another perspective, this time accounting for total headline mentions on each publication’s full homepage.
To some degree, this establishes unfair comparisons. Depending on editorial standards, business models, staff sizes, and other variables, there’s much diversity in how homepages are structured and filled (and by extension, how news is shared). That said, we’ve decided to include total mentions in this analysis—given the gravity of what transpired, any work to inform the public should be noted, and when factual, saluted.
There are many lessons contained within the emergence of this crisis, particularly for those of us who publish information regularly. Regardless of whether it’s in service to journalism or marketing, when we move to produce and publish something, we face similar pressures—to be clear, concise, accurate, responsible, and to make effective use of our narrow opportunities in a wildly competitive attention economy.
We hope all publishers will take this analysis as a cue to reflect on the ways they attempt to capture and direct people’s attention. On most days, the implications may be relatively benign. But sometimes, there will be dire consequences for the way information is disseminated, as there have been over these past several months. In those moments, high and well-practiced standards will benefit us all.
Our analysis was conducted between April 24 – May 18, 2020, sourced exclusively through Internet Archive’s Wayback Machine. Internet Archive is a non-profit digital library founded in 1996 and its Wayback Machine has archived 431 billion web pages since it’s inception.
Did you only count headlines that mention Covid-19 or coronavirus?
No, we did not limit our count to those words alone. Anything that made general or specific reference to the emerging crisis, was counted. We counted headlines mentioning “virus,” “illness,” pandemic” and the like, but also extended into less obvious titles, another reason we couldn’t automate our scans of the pages.
For example, any of these hypothetical headlines would have been counted: “Chinese Government Under Fire For Handling of Crisis,” “Here’s How To Wash Your Hands Correctly,” and “Public Health Officials’ Worst Nightmare Becomes Reality.”
Don’t some websites display more headlines than others? How is it fair to compare two publications with different sized homepages?
Yes, some publications display more headlines on their homepage than others. We addressed this in the analysis above, and to summarize, we chose to compare diverse homepages because we consider the raw attempt to share information with the public most important.
Why didn’t you use every snapshot for a given website in the Internet Archive, during the period you analyzed?
We certainly would have liked to. Unfortunately, these counts needed to be conducted manually. Attempting to cover that much ground would have increased an already very significant research burden well beyond our capacity. Instead, we selected a repeatable sample—Monday afternoons—to establish trends in coverage and to fairly compare results between publications.
Why were certain media outlets left out?
Some websites, intentionally or not, restrict Internet Archive from accessing and archiving their contents. For this analysis, we thankfully had access to a vast majority of the sites we were interested in. Notable exceptions to that are CNN and Associated Press.
For media inquiries, contact firstname.lastname@example.org