Having enjoyed spikes in online traffic and subscription sales in March, news publishers brace for the end of the bump. Is it going to be quick or slow? Will there be more peaks?

The analysis of Google Trends data, a proxy for the audience’s interest due the large volume of data, confirms the COVID-19 pandemic is an unprecedented event in terms of scale and length.

• The demand for online news worldwide doubled versus the averages for the past five years and more than tripled versus the averages of 2005-2010.
• The demand spiked in March and dropped since then. In early April, it still was 56% up vs. the average for the past year.
• In China, the demand peaked at the end of January, then it fell, and then it climbed again. The average for March was 50% up and in the early April it was 58%.
• Thinking of the duration of the spike in demand, we do know that the interest wanes over time, but we do not know the exact duration and the shape of the curve. 

How quickly is the bump going to end? 

In the last edition of the newsletter we described the “hard landing” pattern observed by Mather Economics around high-interest news events. During natural disasters, elections, crimes, or dramatic sports events such as the NFL Super Bowl, the acquisition of new subscribers sped. After the interest waned, the sales suddenly dived, even below the pre-event level.

Some INMA members expressed doubts whether the pandemic would follow the “hard landing” pattern. Lukas Görög of Die Presse in Austria wrote on the INMA Slack channel: “I don’t think the drop will be so dramatic and steep, as in the case of Super Bowl, and I also think it will last longer.” Last week, Die Presse was converting on average three times more subscribers than before the pandemic.

So, what kind of a bump is the right benchmark here? 

The 2019 study on the lifespan of news stories found an average news cycle in the United States lasted only for a median of seven days, with the longest lasting event getting interest for 43 days.

Researchers from Google Trends, Schema Design, and Axios saw the duration of the bumps depending on the speed at which events developed, and whether or not their outcomes were expected, like in the case of elections or the sports game finale.

They identified four patterns:

• Symmetric: steady climb, steady fall. This pattern reflected news events that were anticipated ahead of time and continued to capture interest afterwards, such as elections, new bills.
• Skewed left: slow climb, quick fall. This was an anticipated or ongoing event that ended with a simple conclusion, such as natural phenomena or sports games.
• Skewed right: quick climb, slow fall. This reflected sudden and unexpected news events, such as tragedies and natural catastrophes.
• Non-symmetric: multiple peaks. This was an important event that spawned a secondary reaction or addition, such as trials or scandals with long repercussions, or long crises such as immigration. 

A long crisis with peaks along the way

Thinking of the COVID-19 pandemic, news publishers may expect the demand for the coverage is going to follow a pattern of a long crisis with a high potential for multiple peaks, depending on the development of the outbreak and the lockdowns.

My reasoning is based on the two observations:

• My analysis of Google Trends suggests the demand for news in China has lasted for more than 100 days already, and the second bump can be observed in March, as the new coronavirus cases were confirmed.
• The bump in Italy, the first country in Europe affected by the pandemic, has lasted for more than 50 days so far, and in early April still was 50% vs. pre-Covid averages.

Forecasting future traffic to news sites, publishers may wish to factor the difference in demand to general news and to specialist news categories and other verticals.

• Based on the traffic trends in Europe, where the pandemic broke a few weeks earlier than in the United States, one can expect a shift of attention from public health-related news to economy news as people become more concerned with the state of the economy. 
• One can also safely predict a prolonged drop in attention to some lifestyle categories crippled by the lockdown, such as automotive, travel, etc.

Forecasting future demand for news subscriptions, the core analysis of Mather Economics and other experts stands, in my opinion:

• The COVID-19 spike in sales most likely came from the mobilisation of already engaged readers, and not new ones, as the analysis of another expert vendor, Deep.BI, showed.
• The high-interest news event sped the conversion of engaged readers, but that created a gap in the bottom of the funnel. The new ones attracted to the site for the first time might convert, too, but not in the near term.
• So, the duration of the subscription bump may follow the bump in traffic, only if publishers successfully engage new readers at the top of the funnel and fill the bottom of the funnel over the next weeks and months. 

The research is changing weekly, as it seemingly everything in the world. Subscribe to my newsletter to get my up-to-the-minute research reports.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .