Retailers Face a Data Deficit in the Wake of the Pandemic

by Angel Evan and Amber Rivera - June 22, 2020

When stay-at-home orders took hold this past March, retail sales dropped dramatically - as everybody knows. But that change in customer behavior has resulted in a phenomenon that hasn’t been talked about much: the flow of sales information to retailers’ data repositories has dried up. That’s a significant problem, because a healthy flow of that information is the lifeblood of customer loyalty programs, AI-driven product recommendations, and a wide array of critical business decisions.

What this change means is that many retailers – independent or chain, brick-and-mortar or e-commerce, startup or legacy – are now facing an information deficit. That’s what happens when the data and intelligence derived from customer transactions becomes scarce or unusable due to a sudden change in buyer behavior. Today the problem is widespread: Even businesses that had amassed great volumes of customer data before Covid-19 are finding themselves in the same cold-start position as businesses venturing into unknown markets or reaching out to new audiences.

The effects of this disruption could be massive in the mid- to long term, because it obviously makes customer behavior more challenging to interpret, to predict, and to pattern. In the current context, this much is clear: Businesses should not take for granted that the data they gathered before Covid-19 will accurately predict buyer behavior in the socially distant economy.

Instead, retailers must take careful stock of the data inputs and analytical assumptions that now drive their products and business decision-making. They must identify the risks of carrying on with the status quo, and they must respond to the challenges of the moment with creativity and innovation. This recalibration exercise will help retailers quickly figure out how to stay relevant as American consumers change.

Risks and Opportunities of an Information Deficit

In our post-stay-at-home reality, companies need to recognize that their existing predictive models, forecasts, and dashboards may all be unreliable, or even obsolete, and that their analytic tools need recalibrating. Although the objectives of a specific automated system or predictive model may not have changed, the input data and the users certainly have, which should cause companies to re-evaluate how outputs are interpreted and relied upon.

Companies also need to avoid making shortsighted decisions about data infrastructure and human resources. Although staff reductions can help to offset immediate profit loss, eliminating the people who know how to organize, clean, mine, and model customer data can create intractable technical debt. When making resource decisions, leaders must truly understand which systems will suffer as a result of losing a specific role, then quantify and weigh the longer-term cost of any fallout.

The changes we’ve recently witnessed in customer behavior have left some companies feeling as though they’ve been dropped in the middle of the woods without a compass. The classic “Who is our customer?” question has suddenly become harder than ever to answer. Are current online shoppers repeat customers who migrated from brick-and-mortar stores, or are they net new customers? Retail businesses have always had a blind spot in this area, but until now they haven’t had much reason to prioritize it. Those who make the most of the current stay-at-home moment to gain clarity here could make transformative discoveries that will serve them well beyond 2020.

Which brings us to some good news: This situation represents an opportunity for companies that have not ever fully harnessed data to make evidence-based business decisions. Before the pandemic struck, these companies lagged behind data-mature companies, but as data-mature companies are now struggling to contend with a sudden information deficit, the less data-mature companies have a one-time opportunity to build their data-gathering capabilities and enhance their data-driven decision-making. For a brief period, at least, they’ve got a chance to catch up.

In this moment, it’s important for all companies, no matter how data-mature, to remember that “customer data” is not limited to point-of-sale transactions. Retailers should consider data to be any information that is relevant to their customers’ behavior that can be ethically collected, organized, and studied for insights that decision-makers can rely on. As restrictions on public life continue to ease and shopping resumes, retailers must get creative about where and how to collect this information.

They should, for example, broaden their understanding of shopper behavior to include anything that signals the way a customer relates to a company and its products. There’s lots more companies should do: They should study which messages resonate with different customer segments. They should measure their expectations for fulfillment. They should characterize the nuanced ways customers’ purchasing patterns have been transformed. Companies must not let a temporary reduction in transactions or three months of distorted data impact their ability to make informed decisions. Using the analytical methods of data science, they have to weave many threads together to fashion business resilience in the face of uncertainty.

Data-Science Countermeasures

To defend against an information deficit, it’s essential to understand what relevant data are available and how those data can be reengineered to answer contemporary questions. Companies can examine their own existing data within a smaller timeframe or even design a survey to get a fresh pulse on customers’ needs and plans. Moreover, refreshing machine-learning models and analytic tools will help companies get back on track and begin competing anew.

Here are some ideas available to retailers to get in front of the information deficit problem.

1. Stand up new channels of communication and data collection with the guiding question, “What are we building with our customers while we can’t build revenue?” With thoughtful communication, companies can gain customer respect for their resilience and generate useful data. Learn which messages resonate, what products customers still consider essential, and how much appetite they have for socially distant shopping. Business opportunities lie in the nuances of these data points from individual to individual.
2. Even if stores are closed, or sales are depressed, leverage “Covid-19–aware” data sources that are capturing consumer behavior indicators. This will help you begin characterizing new preferences and purchasing patterns. For instance, there is valuable information to be gained from analyzing email interactions, customer-care call logs, website sessions, and social-media data. As a proxy, study sales data from 2008 and 2009 to understand how customer behavior changes during times of economic hardship.
3. Revisit revenue projections to account for the new constraints imposed on shopping due to Covid-19. Incorporating this information into analyses will support updates to purchasing, staffing, and other important cash-flow decisions. Some variables to consider include historic data about customer traffic, the size of a store in square feet, and adjustment factors to account for a depressed economy.
4. If stores are located in multiple states, use spatial analysis in conjunction with relevant Covid-19 health ordinances and regulations to analyze where it makes the most sense to focus attention upon reopening. Not all state and local economies will reopen to the same degree, and those constraints need to be considered along with more-traditional inputs such as customer demographics, human mobility, and historic store performance.
5. Take this opportunity to give attention to the infrastructure supporting key data assets. This might involve the following: redesigning approaches to data collection and storage such that newly relevant data can be quickly mined for insights; creating or reengineering predictive models using more-focused data sets; fixing glitches in website analytics and tagging practices that hinder the ability to draw accurate conclusions from website data; and revisiting key performance indicators and scrutinizing each formula’s variables. This will ensure that prior assumptions are still applicable and not artificially skewing KPIs.

Covid-19 will eventually fade away, but that does not mean business decision-making should resume its prior incarnation. Data and quality analysis will remain essential instruments for making wise decisions about customers. Retailers that can harness data quickly and smartly will maintain their business and a competitive advantage in the long run.

Angel Evan is the President of Particle Inc. and an instructor in the Continuing Studies program at Stanford University.

Amber Rivera is a Data Scientist at Particle Inc. From start-ups to multinationals, she leads executives to better decision-making through data.