The digital footprint of small businesses and metropolitan resilience during COVID-19
///

The digital footprint of small businesses and metropolitan resilience during COVID-19

How digital strategies, particularly the use of websites by small businesses, influenced the economic resilience of metropolitan areas during the COVID-19 pandemic.

Businesses turned to technology to sustain operations during the COVID-19 shutdowns and subsequent social distancing measures. This study examines how the commercial use of websites impacted the economic resilience of communities. Did this strategy help metropolitan regions and small businesses mitigate the crisis’ effects?

The study focuses on the 50 largest metropolitan areas in the US, which are critical to the national economy and home to 53% of the population. We examine the impact on these communities and their businesses between April and December 2020 as the pandemic unfolded.

While many traditional economic data sources are typically available annually, our study leverages new data collection methods that emerged during the pandemic. This includes data drawn from commercial sources like credit card companies and the introduction of rapid “pulse surveys” from the US census. We use these sources to monitor real-time outcomes, complemented by monthly data covering over 20 million web domains. Domains serve as the foundational address book of the internet. Most domains host websites and support digital functionalities such as email and encryption. 

We use anonymous customer data from GoDaddy, the largest supplier of domain names, to track developments during this period. At least 75% of these web domains are commercial; 92% support microbusinesses with fewer than 10 employees. This study examines the role of these digitally enabled small businesses in fostering community resilience, which is the ability of local economic systems to absorb shocks and recover from disruptions. 

We briefly discuss how going online could influence business and community resilience and outline the data and methods that formed the basis of our analysis. After presenting our findings, we further explore the implications for public policy and metropolitan economies in the future.

What is the evidence on how technology affected resilience for small businesses and their communities during the pandemic? US metros with higher density of commercial website domains had more spending, revenues, new openings and fewer negative effects for small businesses, suggesting this as a critical area for small business development for the future.

Karen Mossberger

How digital tools help small businesses and communities

Between February and April 2020, the number of businesses nationally dropped by 22%, with small businesses being the most vulnerable. They often lacked the cash reserves to endure the shutdowns or the flexibility to make rapid changes. In addition, they were heavily concentrated in the hardest-hit industries, such as retail, restaurants, and personal services.

A study of German businesses demonstrated that websites helped businesses to adapt during the pandemic by enabling them to communicate changes to customers. Websites enabled brick-and-mortar businesses to advertise modified hours, curbside pickup, delivery options, reopening plans, new menus, health protocols, virtual appointments, online shopping, online classes, and consultations.

Web domains can also facilitate new startups and growth. After the initial shutdowns, new businesses started growing by 24% in 2020, with most being online-only ventures and solo entrepreneurs. Websites make it possible to launch a business from home, eliminating costs like rent and utilities. Post-pandemic, brick-and-mortar businesses may have expanded their customer base through online sales, reaching new markets

Apart from the shutdowns, prior evidence indicates that higher domain density correlates with several positive economic outcomes at the county level. These include growth in median income, recovery from the 2008 recession, and overall economic prosperity. Counties with a higher domain density also tend to exhibit stronger labour market outcomes.

The impact of the COVID shutdowns varied by region and was influenced by the mix of industries in the local economy. Therefore, it is important to understand the effects of digitally enabled business strategies across major metropolitan areas. In our analysis, we examine monthly fluctuations in the density of domain sites across the 50 largest metropolitan areas in 2020, exploring their implications for economic outcomes and small business perceptions.

Data on 20 million+ domains

To measure the prevalence of website usage in metropolitan areas, we use customer data on over 20 million US domain sites from GoDaddy, the world’s largest registrar of domain names. The data is anonymised, aggregated by zip code, and available monthly. GoDaddy’s dataset incorporates “redirects”, where multiple websites link to a central domain, consolidating them into a single “venture” to prevent double-counting. Zip-code level data on domains is aggregated by metropolitan statistical areas (MSAs) and divided by the MSA population to create a density metric— the number of domains per 100 residents in the metro.

Although these domains serve diverse purposes, extensive customer surveys and machine learning analyses confirm that at least 75% are commercial. Surveys indicate that more than half of these commercial sites are managed by solo entrepreneurs. GoDaddy’s dataset includes websites for both brick-and-mortar and online-only businesses.

Credit. Midjourney

Measuring change across metros monthly during the crisis

To ensure alignment between data availability and measurement of outcomes and perceptions, we focus on the 50 largest metropolitan areas. For assessing outcomes, we use the Opportunity Insights Economic Tracker developed by Chetty and colleagues at Harvard University. Throughout the pandemic, this tracker has provided insights into trends across metropolitan areas using anonymised data from private firms, including credit card and payroll companies.  Our analysis incorporates metrics such as credit and debit card spending (from Affinity Solutions) and data on business revenues, openings, and closures (from Womply). We calculate monthly mean values to correspond with the frequency of domain data collection.

The time series analysis begins in April 2020, coinciding with the peak of shutdowns and heightened unemployment, and ends in December 2020. Our study focuses on three main outcomes:

  1. Monthly mean values for credit and debit card spending, seasonally adjusted, compared with January 2020 levels before the shutdowns.
  2. Percentage change in seasonally adjusted small business revenues indexed to January 4-31, 2020. 
  3. Percentage change in seasonally adjusted small business openings and closings indexed to January 4-31, 2020.

Examining perceptions reported by small businesses

We also examine the relationship between domain density and the perceptions reported by small businesses, using data from the Small Business Pulse Survey conducted by the US Census. To monitor developments during the pandemic, the census initiated comprehensive, high-quality random sample surveys of small businesses weekly starting from April 2020, with data available for the 50 most populous metros. 

We use two questions from the Small Business Pulse Surveys to portray how businesses perceived conditions, calculating mean values for the month for each metro. The first question assesses: “Overall, how has this business been affected by the COVID-19 pandemic?” We analyse the probability of experiencing a negative or no change based on their responses. The second question asks: “In the last week, did this business have a change in operating revenues/sales/receipts, not including any financial assistance or loans?” Here, we focus on perceived decreases or no change.

We analyse these outcomes or perceptions using models considering changes over the entire period and month-to-month variations. Our analysis controls for several factors that could influence outcomes and perceptions. These include broadband subscriptions in the region, which serve as a measure of local technology adoption. We also account for demographic variables, including education (known to impact metropolitan resilience in previous economic downturns) and considerations of race and ethnicity (given that businesses owned by people of colour have historically faced higher vulnerabilities).

Controlling for additional factors

We also control for relative prosperity before the disruption and account for variations in local industries, which significantly influenced economic outcomes during the pandemic. Regions reliant on tourism and service sectors were particularly disadvantaged. Although certain “superstar cities” with a strong IT presence and knowledge-based industries saw a decline in population during this period, there was a notable shift towards less densely populated areas within the same metropolitan region, largely due to reduced commuting patterns.  

We expect that beyond these other factors, metropolitan areas with higher domain density will exhibit outcomes including increased levels of credit and debit card spending, higher revenues for small businesses, a greater number of small business openings, and fewer instances of small businesses reporting decreased revenues. Additionally, we expect that regions with higher domain density will reflect more positive perceptions among small businesses in the pulse survey, with fewer reports of revenue declines and overall stability in business conditions. 

Evidence for resilience across multiple measures

Analysing credit card transactions, payroll data, and other commercial indicators from Opportunity Insights, we find that metropolitan areas with a higher density of domains demonstrated greater economic resilience from April to December 2020. This resilience was evident in several key metrics, including increased credit and debit card spending, higher levels of small business revenue, a greater number of small business openings, and fewer instances of small businesses reporting revenue declines. 

The analysis using month-to-month changes provides a robust assessment of the effects of domain density on the regional economy over a shorter timeframe, considering the cumulative impact of prior changes. Domain density consistently correlated with resilience across all measures throughout the overall period and for nearly all month-to-month comparisons. The exception was small business revenue decreases, where the results did not show statistically significant differences when examined monthly.

The US Census survey echoes similar findings reported by small business owners themselves. In metropolitan areas with a higher density of domains, small businesses were more likely to report minimal changes in revenue, and fewer reported negative effects during the pandemic shutdowns. While domain density did not yield statistically significant differences in reporting “little or no effect” or in month-to-month analyses, the overall results from the census survey align with the conclusions drawn from commercial data provided by Opportunity Insights. In both cases, we see significant effects for resilient outcomes in metropolitan regions with higher density of domains, even when accounting for other local economic factors such as demographics and industries.

 OverallMonth over month
Have higher levels of credit/debit card spending
Have higher levels of small business revenue
Having a lower level of small business revenue decreases 
Have a lower level of small business revenue decreases na
Have a higher level of small businesses reporting little to no change in revenuena
Have a lower level of small businesses reporting a negative effectna
Have a higher level of small businesses reporting little to no effectnana
Table 1. Communities with higher digital economic activity

Thriving in the digital economy in the pandemic era

The internet provided a lifeline for many businesses during the pandemic, and innovative data collected during this period vividly illustrates its critical role in bolstering the resilience of small businesses and communities across metropolitan regions. Small entrepreneurs were better able to sustain revenues and mitigate the negative effects of the shutdowns in areas with a higher density of domains—a marker of widespread internet usage, particularly among microbusinesses. 

There are, however, significant implications for the economy and public policy that extend beyond the pandemic shutdowns. Previous research conducted before COVID-19 indicated that domain density was a significant predictor of economic prosperity in US countries. The pandemic merely accelerated digital adoption among small entrepreneurs, leading to the rapid growth of new online small businesses. Small business programmes and local economic development policies should consider how to support digital skills for brick-and-mortar businesses. It must also recognise the role played by small online entrepreneurs in regional economies. 

With historic investments in broadband and digital equity, there is an opportunity to foster digital skills and cultivate the growth of digital businesses in both urban and rural communities. 

Finally, the urgency of the pandemic crisis inspired new efforts to measure developments in real-time. This includes gathering commercial data and conducting high-frequency surveys through the census. These shifts in data availability from diverse sources hold promise for enriching our future understanding of local economies and their development. 

This evolution in data collection methods could lead to more informed decision-making and better-targeted policy interventions to support economic resilience and growth across metropolitan areas and beyond.

🔬🧫🧪🔍🤓👩‍🔬🦠🔭📚

Journal reference

Mossberger, K., Martini, N. F., McCullough, M., & Tolbert, C. J. (2023). Digital economic activity and resilience for metros and small businesses during Covid-19. Small Business Economics60(4), 1699-1717. https://doi.org/10.1007/s11187-022-00674-x

Karen Mossberger is an emeritus professor in the School of Public Affairs at Arizona State University and former director of the Centre on Technology, Data, and Society. She researches technology use and its impacts, digital inequality, urban policy, and local government. She is co-author (with C. Tolbert and S. LaCombe) of "Choosing the Future: Technology and Opportunity in Communities" (Oxford, 2021). Karen Mossberger is a fellow of the National Academy of Public Administration and the recipient of the 2023 NASPAA/ASPA Distinguished Research Award.

Nicholas F. Martini is a Lecturer and Director of Undergraduate Studies in the Department of Political Science at the University of Iowa. His research focuses on the intersection of international relations and political behaviour. His current research explores the factors driving public opinion (e.g. ideology, beliefs, and religion) and how they shape preferences on foreign policy issues.

Meredith McCullough is a PhD candidate at Arizona State University’s School of Public Affairs and a research associate at ASU's Centre on Technology, Data, and Society. McCullough's research focuses on local economic development policy, entrepreneurship, and the dynamics of digital transformation.

Caroline J. Tolbert is Distinguished University Professor of Political Science at the University of Iowa, where she teaches graduate seminars in statistics and undergraduate courses in public policy and social media in politics. She was named a 2021 Andrew Carnegie Fellow for her research on voting and state election laws. She is co-author of "Choosing the Future: Technology and Opportunity in Communities" (2021), which won the 2022 Goldsmith Prize for the best book from the Shorenstein Center at Harvard’s Kennedy School of Government.