The 6 Keys to Business Forecasting in the COVID-19 Era
Companies have never had to deal with a crisis like the coronavirus recession. Deciding when and how to reopen businesses after the first big wave was only the beginning; managers now are trying to solve a jigsaw puzzle of subsequent waves of infections and local lockdowns. While approved vaccines are providing hope, their slow rollout is a painful reminder that dark days lie ahead.
Yet while COVID-19 impacts almost all aspects of the economy, most businesses still don’t have a solid understanding of the underlying forces driving the spread of the virus or how to model for them. Even now, months into the pandemic, executive committees in a range of industries have been forced to rethink return-to-office schedules, re-assess demand predictions, redraw their supply chains or revise their financial targets (again).
The key to business forecasting during these unprecedented times is to factor in the six most important drivers of COVID-19 transmission. Four of them are epidemiological: undetected cases, “nonstationary” human behavior and vaccination roll-out, “heterogeneity,” and the mixing of people throughout society. The other two are purely statistical: mathematical assumptions on various unknowns and, yes, general randomness.
Taken separately, each of these factors can help businesses improve their planning to some extent. Those that are able to account for all six are developing much sharper forecasts and smarter scenario analyses that set them up to better navigate through the coronavirus recession.
The first step in forecasting is factoring in undetected COVID-19 cases. Medical studies show there is a much larger volume of undetected cases than detected ones, whether they be asymptomatic infections, false negatives or insufficient testing. The ratio of undetected to detected cases differs by region and over time. The U.S. average is currently 1.8 to 1, according to Oliver Wyman estimates. But while Miami has close to double the infection rate per capita compared with New York City based on detected cases, Oliver Wyman estimates that New York City has more than twice the infection rate per capita based on the total of detected and undetected cases. So to focus only on the number of reported cases is to see only the tip of the iceberg.
Instead, businesses should monitor the estimates for undetected cases so they can recognize when immunity levels rise in certain regions, and therefore become more comfortable with bringing people back to offices. Some financial services firms in New York, where COVID-19 transmission and undetected cases are way down from the spring peak, have been able to bring back a larger number of employees to the office based on this type of immunity analysis.
‘Nonstationary’ Human Behavior
The next step is to consider how human behavior is changing in response to the virus and how the number of people still susceptible to the disease is declining with vaccinations. When COVID-19 began to spread, individuals made substantial changes to their routines, such as avoiding crowds, restaurants and gyms and using personal protective equipment. These actions were “nonstationary,” meaning they changed over time and are likely to continue to change in the future. The behavioral responses aren’t uniform — masks are being worn religiously in many regions, while residents of some areas are burning them in protest. Bars are packed with customers one day and light the next.
Taking these nonstationary behaviors into account can give companies a major leg up in planning, seizing opportunities and controlling costs. Several retailers have developed more accurate demand and store-traffic predictions by monitoring changes in mobility. This has allowed them to right-size their staffing to accommodate that traffic, managing costs and improving profitability.
Finally, as more individuals become immune (if only temporarily) to the virus through vaccinations and infection, the combined effect creates more protection against virus transmission throughout the society.
Business leaders who can consider critical factors of return-to-work separately and together will be better equipped to navigate the worst pandemic in a century.
Businesses also should consider the ways that groups of people of different ages and in different settings such as work, home and school interact. Some people commute to work using public transportation, while others are retired and stay mostly at home — and COVID-19 spread rates vary greatly among those groups. In the United States, the CDC has outlined a sequential vaccination strategy starting with health care workers and the most vulnerable.
Such differences in behavior naturally slow the spread of COVID-19 — the most active and least cautious individuals in society get the virus early and become immune, while those who are less active and more cautious are less likely to contract the virus. This analysis helps pharmaceutical firms in vaccination plans: They plan to vaccinate frontline health care workers first, because they are the most exposed to the virus and interact the most across different groups of society. At the same time, providing immunity to the most vulnerable will reduce the strain on our health care system and, most important, reduce the death toll.
Mixing in Society
Another related step in planning during the pandemic is to understand the phenomenon of mixing in society. A big outbreak starts with smaller clusters before it spreads broadly. Colleges in the United States, for example, experienced numerous clusters last fall as students returned to school. The extent to which those students interacted with the surrounding towns — say, going to restaurants or doctors’ offices — determine the severity of the outbreaks and the damage they do to the local economy.
Businesses that recognize the negative potential of mixing are springing into action when a cluster develops to ensure that mixing remains minimal and their operations remain as robust as possible. One South Korean company, for example, successfully contained a coronavirus outbreak by identifying the person at the center, testing everyone who came into contact with the worker and imposing tighter quarantines. Reducing mixing helped limit the spread of the virus both throughout the workforce and beyond, allowing the business to remain open during the pandemic.
The final two steps in forecasting stem from the world of statistics. The first is accounting for unknowns: Are there factors for which no one can reasonably estimate the outcome? Businesses that control for them by simulating a wide range of assumptions are far more successful in planning for the future.
The massive unknown of how many people will choose to use approved vaccines affects demand scenarios for companies across virtually every economic sector, since vaccinated people will feel more comfortable participating in all aspects of the economy. Pharmaceutical firms have begun projecting wide ranges of COVID-19 spread for the next 10 years along with vaccine coverage and efficacy assumptions. That’s allowing them to better predict demand for their products and formulate their financial, production and distribution plans.
Finally, companies have to assess the flukiest factor of all: pure randomness. Some events simply aren’t predictable, and that volatility is especially critical in modeling COVID-19 spread. One unexpected random event can have a massively outsized effect. Forecasts that incorporate the chance of randomness are better able to understand the scope of the problems and solutions.
For example, hospital managers have been seeing sudden, random surges in demand for their non-COVID-19 services based on volatile COVID-19 patterns. One consequence of that randomness is that they are fielding many more digital appointments than ever before. To plan for future deluges, many are now building in more digital capabilities and considering other ways of doing business that will endure long after the COVID-19 crisis.
Modeling all these critical factors is no easy task. But business leaders who can consider them separately and together will be better equipped to navigate the worst pandemic in a century.