Machine Learning Model Uses Advanced Analytics to Forecast Consumer Behavior across 10 Categories and 15 Countries

First Report on Automotive, Insurance and Food Delivery Suggests Greater Rebound in U.S. vs U.K. across Adversely Impacted Categories

UM, the global marketing and media agency network of IPG Mediabrands, released the Consumer Demand Forecaster, its proprietary machine learning model that predicts the rise and fall of consumer demand in various markets amidst the COVID-19 crisis. The accompanying series of reports provides a roadmap for brand marketing re-entry strategies across 10 major categories, including Automotive, Consumer Electronics, CPG, Entertainment, Finance, Food, Insurance, Oil & Gas, Retail and Travel.

The first set of findings predicts consumer trends across highly impacted categories including Automotive, Insurance (automotive, motorcycle, home, life, travel and cyber) and Food Delivery for the U.S. and U.K. markets. In the Automotive and Insurance categories, there are consistent signs that consumer demand in the U.S. will rebound in mid-May and approach pre-COVID-19 levels by early June. In the U.K., however, there are little to no signs of recovery in consumer demand over the same period.

Marketing Technology News: Vestorly’s Software Powers FMG Suite’s AI-Driven Content Engine

Using online search data as a representation of consumer demand, the Consumer Demand Forecaster leverages advanced analytics to predict the timing, rate and extent of recovery in consumer demand across 15 countries in different stages of the COVID-19 crisis, enabling marketers to plan their re-entry appropriately. Additional insights across categories will be released weekly.

“Our data-fueled model provides practical, actionable media guidance to help brands respond to and prepare for recovery from the COVID-19 pandemic,” said Huw Griffiths, Global Chief Product Officer, UM. “By predicting the rise and fall of consumer demand across categories, and the magnitude and shape of that demand, we are able to advise brands on when and how to re-enter the market in both the short- and long-term, with the goal of future-proofing their businesses during this uncertain time.”

The model tracks a variety of key factors that influence consumer behavior in each category, including: new, confirmed and recovered COVID-19 cases; days since 100 confirmed COVID-19 cases; the consumer confidence index (CCI); the number of government restrictions; household size; disposable income; unemployment; seasonality and more.

Marketing Technology News: Fjorge and B2B Ecommerce Platform Znode Announce New Technology Partnership

The initial findings from the Consumer Demand Forecaster provide predictions across three key categories for the U.S. and U.K. that have experienced significant shifts in consumer demand amidst the COVID-19 crisis: Automotive, Insurance and Food Delivery.

Griffiths continued: “Our model suggests that recovery is expected to occur more quickly in the U.S. than in the U.K. across Automotive and Insurance, which could be attributable to the U.K. experiencing its biggest drop in consumer confidence on record as a result of the Covid-19 crisis.”

The Automotive category experienced an immediate and significant decline in consumer demand between March and April, with a greater chance of recovery in the U.S. in May.

• Consumer demand for automobiles in the U.S. will begin to recover in early May, with a marked increase halfway through the month.
• U.S. demand is predicted to rise to 80 percent of pre-pandemic levels by the end of May, suggesting a U-shaped recovery curve.
• Consumer demand for automobiles in the U.K., however, is expected to show little increase in May, stabilizing at 60 percent of pre-pandemic levels and may stay flat or fall slightly by the end of the month, suggesting an L-shaped recovery curve.

While the Insurance category experienced an initial spike in consumer demand, there was a significant decline in March with some signs of a U.S. rebound by the end of May.

• Both the U.S. and U.K. markets saw a spike in consumer demand as the COVID-19 pandemic hit, but the spike in the U.K. was larger and longer than in the U.S.
• Signs of a rebound in the U.S. begin in mid-May and point to a return to near pre-pandemic levels by the end of May, suggesting a U-shaped recovery curve.
• Consumer demand for insurance is expected to stabilize in the U.K. through May at 10-15 percent below pre-pandemic levels.

While the Automotive and Insurance categories have suffered during the pandemic, demand for Food Delivery has experienced a significant increase, and this demand will likely remain higher than pre-pandemic levels in both the U.S. and the U.K. throughout May.

• In general, Americans ordered food delivery more than their British counterparts prior to COVID-19, but both markets saw nearly a doubling in demand from mid- to late-March.
• The model predicts that demand for food delivery in both markets will remain at elevated levels through the end of May, with the U.S. potentially seeing a small dip in the last part of the month while the U.K. remains relatively stable at the elevated levels.

Marketing Technology News: Sharon Loeb Joins Ellucian as Chief Marketing Officer

The Consumer Demand Forecaster is UM’s proprietary model that uses global data, advanced analytics and machine learning to quantify and predict the impact of COVID-19 on consumer demand. The machine learning technique applies a regression tree methodology to global COVID-19 data that is then analyzed to build predictive models to explain how consumer demand, expressed by search data, responds to various factors, including macro-economic and structural variables such as consumer confidence, income dispersion and seasonality. The modeling was conducted across 15 markets at differing points on the COVID-19 Response and Recovery curve, including Australia, Canada, Denmark, France, Germany, Hong Kong, Iran, Italy, Japan, Mexico, Singapore, South Korea, Spain, the U.K., and the U.S. The Consumer Demand Forecaster model is updated weekly with the latest data as well as the latest epidemiological forecasts on COVID-19 projections from the Institute for Health Metrics and Evaluation (IHME), an independent global health research center at the University of Washington; and with deeper dives into each market, as well as a state-by-state breakdown in the U.S., to keep marketers informed as the COVID-19 situation evolves.

“Our model leverages deep datasets and mathematically-quantified consumer trends across markets that are in the various stages of COVID-19 escalation, response and recovery,” said Tara Connington, Senior Vice President, Analytics & Insights, UM. “This data-first approach enables each country to apply the model in their individual markets to create tailored response and recovery strategies for local brands.”

Marketing Technology News: New Data From Kustomer Reinforces The Critical Role Of Customer Service During The COVID-19 Crisis