How do we quantify the impact of Coronavirus? We used Google Search data to benchmark the reactivity of different industries and mapped the impact by country and industry.
Without repeating what I’m sure you’ve heard already, Coronavirus is having an impact on our lives.
In this report, I offer some of our latest insights into how the COVID-19 situation is impacting many different industries around the globe through search data.
Specifically, it is an attempt to quantify the impact on consumer interest by looking at actual search requests made on Google as well as the level of confidence investors have in the economic viability of particular industries by looking at market returns.
What sets this analysis apart from similar pieces of insight is that we’ve taken a strongly educated guess at what 2020 might have looked like without coronavirus in order to benchmark the reactivity of different industries.
This analysis has context by asking “how much better or worse are industries doing in comparison to where they should be?” rather than “how are they performing?”
What follows is a guide to help you either look at the impact as a whole or drill into more specific and niche industries.
This analysis is intended to provide business leaders, analysts and anybody else interested with new data and insight into the implications the situation has on their clients or business. Search interest in industry adjacent spaces may provide actionable business intelligence for your company, such as subcategories within your industry that are actually performing quite well amid the crisis that you could potentially diversify into.
What is presented in this post are only highlights of this analysis. We have produced a media pack containing all of the charts (as images) as well as the raw and transformed data used to produce them. If you would like to be sent this, please email us at discover@found.co.uk.
Google is a part of most peoples’ daily lives. It is the world’s leading and most popular search engine with almost 2 billion users worldwide and therefore clearly has a keen insight into what people around the world are interested in. Just as interesting, is Google’s data collection of the tens-of-thousands of search queries made every second, which allows users to view and analyse this useful data for marketing and research purposes. This is why we have chosen to use Google Trends to analyse how interest in various topics has been impacted by the COVID-19 outbreak.
Google Trends can be an extremely useful tool for reviewing a company’s digital marketing strategy, especially when we mine the raw data to predict the next wave of popularity or, in this case, paint a picture of what 2020 was supposed to look like. It enabled us to look back across history at how the popularity of any given topic had changed, which is how we have used it in this analysis.
We downloaded the last five years of “interest” data across 274 different industries in 14 different countries with the help of the Google Trends API. The first step was to flatten out any seasonal effects in the data (e.g. “Christmas trees near me” spiking at Christmas). This is standard econometric practice used when forecasting a time series. It removes seasonal patterns underlying the raw data allowing us to look at the trend and other unexplained fluctuations in isolation from seasonal shifts. For example, this graph shows how the highly repeated seasonal pattern (where it reliably peaks around the summer and dips around Christmas every year) in the interest in “Hotels & Accommodation” across the UK can be leveled out to expose the raw trend and other unexplained movements.
We trained a predictive model for each category-country combination based on all data from 2015 to 2019. The models were then used to take a “best guess” at what 2020 should look like based on historic performance. This is our interest benchmark.
Modelling the data gives us a view of what 2020 would have looked like, based on 5 years of prior data. We can get an idea of how different topics have been impacted in reaction to the coronavirus outbreak by comparing actual topic interest in 2020 against what we expected to happen. If a topic drops significantly below our benchmark then we can say that interest has definitely fallen.
Below is a heat map array of how each industry has been impacted in each country we looked at. Some industries have responded really well to the virus outbreak, some are left trying to absorb a massive hit. Many industries are also relatively unaffected. We’ll look in more detail into those soon but let’s start with an overall painting of all topic categories around the world.
We have distilled the above waffle plot into winners and losers to pick out the topics that are thriving and the ones that are receiving the most severe drop-off in interest.
Many topics on this list are a given, such as medical equipment and news. Some are more novel, such as:
Unsurprisingly, restaurants, live sporting events and travel are taking some of the greatest hits. Again, there are some more novel findings here:
We can also drill down into particular overarching topic categories to uncover any aspects of wider interests that are doing well despite the overall trend. For example, let’s look at the shopping industry.
The shopping industry is a very mixed bag. Interest in ticket sales, outdoor markets and luxury items have dropped significantly around the world. But it’s good news for digital marketing as shopping portals and search engines soar in popularity in the UK by 11 interest points compared to where they normally are.
What are they buying? It looks like gifts, consumer electronics, entertainment products and good old ciggies. Auction sites like eBay are particularly popular in South Korea, Italy and Spain. People in Australia are more interested in shopping for jobs over consumer items. Charts for every major category are available in the media pack.
Similarly, we can view a country’s situation at a glance.
In France for example, indoor activities like crafts and digital arts have seen spikes in popularity online but not as profoundly as subcultures & niche interests.
Being on lock down it’s unsurprising how social media and teleconferencing activity has seen a much higher demand. But what is interesting is how people choose to connect online. Virtual worlds have seen the largest jump in popularity in France, rising more than 20 interest points above the expected level. This mirrors the recent rise in online gaming in France as well.
As mentioned, another way to gain insight into how industries are changing and performing during these times is to look at the global stock market. We analysed over 7,000 stocks from around the world in an attempt to understand how investors perceive the economic viability of various industries. We looked at daily returns as a proxy to the value prescribed to different industries by the confidence of investors. Far be it for us to think we can predict how the stock market was going to behave when companies filled with algo traders try and fail to do the same thing; we take this to show “given the last 5 years, what might 2020 have looked like?” Below is a very similar heat map array to the one above for trends, just for stocks. However, we haven’t dived into as much detail here since even 7,000 stocks weren’t sufficient to build a robust portfolio for every industry in every country we were interested in.
At a glance it looks like a lot of bad news. Some noteworthy insights:
If you have any questions about how I analysed the data or if you would like to chat about how we use data here at Found, Get in Touch. Feel free to use any of the assets in the media pack for any stories / content you would like to create off the back of this analysis.
All we ask is that you credit Found by tagging us on social media (links in the footer) and linking back to this page.