The attention that coronavirus (aka COVID-19) is accumulating from journalists is unprecedented — 1.3 million articles and counting. Engagement with COVID-19 content is also without parallel. Audiences are engaged with the media in ways we have not seen before.

Below, our Data Science team shares their analysis of COVID-19 content published between March 9 and March 23. Their research spanned 26 million articles (representing a whopping 82 billion pageviews), including the 1.3 million pieces about COVID-19*.

Coronavirus attention and reader engagement: Our findings

Unsurprisingly, the number of COVID-19 articles began to increase rapidly around March 9, corresponding with the rise in U.S. coronavirus cases and surges across the world that put other countries such as Italy in lockdown. Subsequent coverage on the topic grew as well, with 121% more articles published on March 18 than on March 9.

These findings coincide with the significant jump in cases, and as a result, people seeking the latest coronavirus news. As of the week of March 16, we also found that: 

• COVID-19 articles received about one-third of all pageviews, a 44% increase compared to the week before.
• They account for approximately 35% of Total Engaged Time, up 118% from the prior week, as we see below.

When it comes to the actual amount of time audiences spent actively viewing pages (e.g., clicking and scrolling), we found:

• The Total Engaged Time increase is related to an increase in Average Engaged Time on COVID-19 articles, which now hovers just below 40 seconds. 
• Average engagement across all (non-coronavirus) articles is 35 seconds, as shown in the graph below.

To put these latter points on engagement into perspective: 

• We’ve seen that 45% of readers will leave within the first 15 seconds of reading an article.
• Even greater, 60% of those readers will not return.
• Engaging readers past that 15 second threshold is critical — they’re far more likely to return to your site.

As we’re witnessing, digital content creators have blown that number out of the water.

(For more on Engaged Time, see: Using the Engaged Time metric to grow audiences, unlock loyal readership)

Search and social media trends on coronavirus-related articles

Overall search and social traffic to COVID-19 articles is also up significantly. We found search and social data to be fairly similar, with social driving slightly more traffic when dark social (non-direct traffic that comes to a site by without referrer data) is included, as we see below.

Google Search has shown the most growth across all articles in our network. Even more remarkable — there were more than 500 million daily pageviews from Google Search from March 16 to 19.

When comparing Facebook and Google traffic, we see Facebook hasn’t grown at the same pace as Google in terms of Total Engaged Minutes, as the graph below shows. In fact, the Average Engaged Time from Facebook is down 11% — to 31 seconds on March 18 from 35 seconds on Feb. 25.

We’ll pay close attention to this data, but it’s not wise to count out Facebook’s traffic influence just yet, considering their past recoveries.

Our newest data sent straight to your inbox — subscribe to Chartbeat’s monthly newsletter

Which referrers drove the highest traffic to COVID-19 content?

As of March 18, Google News drove the highest traffic to COVID-19 related content, with Chrome Suggestions not far behind. This was followed by Twitter, Bing, Flipboard, SmartNews and Yahoo!.

From a growth standpoint, most referrers have seen a sharp increase in traffic since March 8. For example, Bing has seen 110% growth, while Twitter and SmartNews have each grown by nearly 60%.

Google News (60%), Flipboard (46%), Yahoo! (35%) and Google Chrome Suggestions (25%) round out the list — and by no means lack in growth.

(Related reading: Mobile aggregators: The next major source of referral traffic)

How does COVID-19 compare to other major events?

When we compare coronavirus reader behaviors to audience interactions during major events, we see many similarities. Here’s what we found:

• Live blogs have by far the highest traffic (41% of traffic to the top 50 pieces), with 66% coming from internal sources (i.e. the homepage)
• However, live blogs underperform on social compared to total traffic, while infographics and live blogs see a below-average proportion of traffic from social. 
• About 38% of articles see more traffic from social media, while interactives and infographics have the highest average search pageviews (approximately 30%).

These findings may sound familiar — they echo past trends we’ve seen across search and social channels during major events.

When we previously analyzed reader trends around 2016 U.S. Election and Brexit, we saw that search and social readers leaned towards certain content attributes. For example, breaking news and “only the facts” types of content saw more referrals via search. In contrast, content that conveyed more emotion, such as Brexit’s “regrets and anger about results” gathered more social momentum.

What can we make of this? 

While COVID-19 is an unprecedented event, content creators still have the chance to establish authority on search and social media. Once readers have clicked through to an article, the key is keeping them on your site. 

Think about good SEO practices (linking to relevant content, research on similar search terms throughout their articles) and social media experimentation (time of posting, hashtags, audience interactions).

(Related: Putting research into play to prepare for 2018 World Cup website traffic)

What is the coronavirus audience data telling us so far?

Coronavirus is unprecedented in both its global imprint and ability to attract sustained attention from readers, if the past month (and arguably since January) is any indication. Publishers, media companies, and content creators should especially take note of the following takeaways:

• Readers are actively seeking information on coronavirus at an increasingly high rate, commensurate with the spike in coverage. 
• This is demonstrated in exponential growth in pageviews, Total Engaged Time and Average Engaged Time across our analysis of 26 million articles representing 82 billion pageviews.
• Live blogs and articles are attracting the greatest levels of attention from audiences, with interactives not far behind
• Growth in coverage and pageviews reflects data patterns we’ve seen in past events such as Brexit and the 2016 U.S. Election. Homepage audiences gravitate towards breaking updates, search readers choose straightforward, personally-relevant information, and social readers spend their time with the human and emotional angles on the news.

Getting readers to your site is not the issue right now, as our analysis shows. Rather, it’s maintaining a high level of engagement with those audiences over time (coronavirus content or not). To ensure that remains the case, we believe digital publishers, media companies, and content creators can all apply optimization tactics that encourage deeper reading.

In the meantime, we’ll be keeping a close eye on these trends as conditions persist across the world.

*Notes from our Data Science team on how we collected COVID-19 data for this article analysis:

We reviewed articles containing the following entities: Coronaviru’, COVID19, Covid19 and keywords ‘%coronavirus%’ OR keywords LIKE ‘%Coronavirus%’ OR keywords LIKE ‘%COVID-19%’ OR keywords LIKE ‘%covid-19%’ OR keywords LIKE ‘%Covid-19%’ OR keywords LIKE ‘%novel coronavirus%’ OR keywords LIKE ‘%corona-virus%’ OR keywords LIKE ‘%Koronaviruss Covid-19%’ OR keywords LIKE ‘%covid 19%’ OR keywords LIKE ‘%COVID%’ OR keywords LIKE ‘%koronaviruss%’ OR keywords LIKE ‘%covid%’ OR keywords LIKE ‘%covid19%’ OR keywords LIKE ‘%koronavirus%’ OR keywords LIKE ‘%corona virus%’ OR keywords LIKE ‘%2019-nCoV%’ OR keywords LIKE ‘%koroonaviirus%’ OR keywords LIKE ‘%COVID 19%’ OR keywords LIKE ‘%Covid-19%’ OR keywords LIKE ‘%Covid19%’ OR keywords LIKE ‘%COVID19%’ OR keywords LIKE ‘%nCoV%’ OR keywords LIKE ‘%Covid 19%’ OR keywords LIKE ‘%CORONAVIRUS%’ OR keywords LIKE ‘%2019 nCoV%’ OR keywords LIKE ‘%2019-nCoV ARD%’ OR keywords LIKE ‘%SARS-CoV-2%’