May 5, 2020
By: Ozgur Dogan, President of Data Science
The world has changed dramatically since COVID entered our lives earlier this year. We have seen travel and entertainment industries take a big hit in terms of revenue and transaction volume while other industries such as online grocery and education sectors have seen significant increases in their sales volume and new customer numbers. As we were observing these macro level market shifts, we were interested in understanding if and how the behavior and performance has changed at individual and geo-levels during the pandemic.
More specifically, we wanted to answer 3 questions:
• Has business performance changed differently across geographies since the COVID pandemic started?
• Is there any relationship between COVID cases and business performance at local market level?
• How has the responder and shopper profile changed during the pandemic?
Our analysis indicated that there is strong relationship between COVID cases and market level performance. We looked at number of COVID cases per 100K population across markets, rank ordered them by incidence rate and created tiers of geographies.
In the financial services area, we saw that the COVID incidence rates had direct relation with performance change from before to after COVID lock down periods. We analyzed new customer acquisition performance for the ‘before’ period defined as February 15th to March 15th and ‘after’ period defined as March 15th to April 15th. We have seen performance declines across all the tiers, but the performance decline in the top tier with the highest incidence rate was significantly larger at 21% vs. the bottom tier which only declined by 3%. The below chart graphically illustrates the relationship between COVID cases and performance changes.
Top tier consisted of hardest hit geographies such as New York, New Jersey, Michigan, Connecticut, and Louisiana. The lowest tier consisted markets such as Oregon, Minnesota, Maine, Arkansas, and Hawaii. While the example shown here involves a performance decline, we have also seen performance increases in other industries.
Separately, we wanted to determine how the responder and shopper profile has changed from before to after lock down. Across many industries we analyzed, we have seen measurable shifts in the responder and shopper profiles. Attached chart shows the changes in the age distribution of responders from before to after lock down. In some industries, we observed that younger, single consumers to become more responsive after COVID, and in others, we have seen older age groups become more responsive and increase their shopping activity. While these specific profile changes cannot be generalized to all industries, recent responders and shoppers after COVID have different profiles. Capturing and incorporating these changes into the targeting models will help increase the marketing performance for addressable marketing channels.
Considering these findings, we see audience targeting change in 2 major ways after COVID:
1. COVID is impacting local markets differently. Our analysis showed strong correlations between area level COVID cases and market level performance.
2. Responder and shopping profiles are changing after COVID. Our prediction is these shifts will continue as the markets go through various stages of recovery and deal with potential new outbreaks in the future. We have seen shifts in the profiles of responders and shoppers for several industries which necessitate the frequent updating of the targeting models based on most recent responder and new customer data.
Implications for Marketers
We all want the COVID pandemic to end quickly, but experts predict it will be here to stay for a while until robust vaccines and medical therapies are developed. The new Coronavirus is likely to keep spreading for at least another 18 months to two years, a team of longstanding pandemic experts from the University of Minnesota predicted in a report released last week. Other expert prediction is we will continue to see a series of smaller waves during summer months to be followed by potentially larger waves in the upcoming fall and winter seasons. Mini regional outbreaks are also predicted which will have impact on small geographies. We are certainly not experts in infectious diseases and don’t have a PhD in molecular biology but it’s wise for all of us to assume that we will need to learn how to live with COVID-19 during next 18 to 24 months.
Given all this, here are our recommendations to marketers seeking performance during this period.