As the economic effects of COVID-19 continue to unfold, the data science team at Monetate analyzed how our clients have been impacted by changing consumer behaviors. For those that aren’t familiar, Monetate delivers data-driven, personalized customer experiences to the e-commerce websites of medium-to-large-sized businesses across a wide variety of industries.
In this post, we’ll take a look at how e-commerce traffic has changed for particular industries over the last two months as public awareness of COVID-19 increased.
To accurately estimate the independent effect of COVID-19, we modeled our clients’ weekly page views, accounting for general trends and seasonal patterns in their site traffic, and included Google search popularity data to quantitatively estimate how COVID-19 awareness approached its saturation point in the general public.
Web Traffic Changes by Industry Vertical
Our modeling allows us to produce statistically significant estimates of the precise effect COVID-19 has had on our clients’ web traffic. As we would expect, the effects varied by industry. Some industries have been impacted more than others, some not at all, and some have even seen an increase in web traffic that can be attributed to the pandemic.
The hardest-hit industries come as no surprise: social distancing behavior has had a significant effect on travel and event service companies. Booking sites (grouped under travel services above), road-side assistance services (same), vacation cruises, ticketing services, and resorts and hotels (grouped under lodging, in pea green) have all experienced a 16-21 percent drop in traffic. Even luggage and bridal retailers saw fewer visitors with an estimated loss of 13-15 percent.
Note: Noticeably absent for our list of industries affected by COVID-19 are apparel and fashion. Although we are beginning to see signs that these industries will be affected by the economic impact of COVID-19, at this point it is too soon to state with confidence what extent the early effect we’re seeing is directly attributable to COVID-19, rather than simply natural variation in the data. As the economic strain continues to be felt around the country, we’ll be monitoring the situation closely and look forward to sharing updated results about the apparel and fashion industry as we have them.
The following graphs show the web traffic these industries experienced in 2020 (red line), along with an estimate of what they would have seen in absence of a pandemic (the dashed blue line), and we also include the prior year’s data for reference (the pink line).
The first week of March was clearly the inflection point in the United States. On February 29th, Washington Gov. Jay Inslee declared a state of emergency. Just under two weeks later, President Trump extended this state of emergency nation-wide. During this period, Americans began to realize the gravity of the situation and changed their consumption behavior to reflect their increased concern about large gatherings and travel.
Our web traffic models allow us to estimate what the level of traffic for these industries would have been in February and March of 2020 without the effect of COVID-19 influencing consumer behavior. More precisely, for each client we fit a linear model that includes terms for significant monthly effects as well as a term for public concern about COVID-19, represented by the popularity of Google searches for “coronavirus.” Errors in this model are represented as a “SARMA process” (Seasonal Autoregressive Moving Average), which accounts for periodic, lag, and spike effects in the time series. Fitting such a model results in an estimate of the independent effect that COVID-19 has had on traffic.
By subtracting this effect out of the time series of true page views, we estimate what the page views would have been without the impact of COVID-19.
As we can see in the plot of average effects above (titled Industries Impacted by COVID-19), some retailers actually experienced a surge in traffic related to COVID-19. We estimate COVID-19 led to a 13 to 15 percent increase in traffic for some retailers in both faith and firearms industries. With students home from school and more people working remotely, home office and children’s arts/crafts supply retailers saw a significant increase in traffic as well, with an estimated percent change of 17 and 28 respectively.
People are also looking to fitness equipment retailers to help them beef up their home gyms. And as the market fluctuates, many are seeking out personal investment resources to stay informed and manage their finances. There was also a very large increase in traffic to retailers of vegetable seeds and gardening tools and supplies. These e-commerce sites saw an average increase in traffic of around 45 percent.
Whether your company has been impacted positively or negatively by the COVID-19 event, there are a variety of ways to optimize your customers’ experience in a data-enriched manner to mitigate the impact of falling demand, or better handle the influx of new visitors to your site. By weaving this data into your onsite experience, you can better act on a customer’s intent, providing better-optimized experiences for optimal engagement and higher conversion during these oscillating times.