Between March and August 2020, one in five consumers switched brands, and seven in ten tried new digital shopping channels. The retail sector experienced ten years of growth in digital penetration in a matter of months. But the resulting surge in data has not provided marketers with substantially better understanding of their customers, because their companies’ outdated data modeling isn’t able to capture these shifts with the necessary granularity and speed.
Rather than using the data to try to better target customers and tailor messages, many marketers have reverted to mass communications and promotions. As one CMO told us, “I’ve largely retreated to mass marketing instead of data-driven marketing because customer behavior is changing so fast I can’t trust my historical data and models.”
But some marketers are accepting the data for the bounty it is and, rather than stepping back from precision marketing, are doubling down. A consumer-goods company, for example, anticipated that sales of beauty products would spike as communities eased out of lockdown. Marketing teams tracked reopenings on a county basis, using epidemiological statistics, municipal reporting, and traffic data to determine where to focus their media spend. These tactics drove a double-digit increase in sales.
Similar insights helped a business service provider get a jump on another emerging trend. Business registration and employment data showed that small healthcare providers in major metropolitan areas were growing at a much faster rate than other small and midsize businesses. Armed with that insight, the company created healthcare-specific product bundles and has launched paid media ads to target those businesses and locales. These moves, combined with other, similarly data-driven campaigns, are poised to increase sales in a core product by more than 10 percent.
Companies that hone their precision marketing in these ways can drive significant customer acquisition during periods of convulsive change. Capturing this opportunity, however, will require brands to update their modeling—from pulling in new sorts of data to retraining algorithms—in order to both keep pace with changing needs and expectations as well as anticipate shifts in customer behavior.
New challenges to account for
Precision-marketing models are trained to recognize and draw inferences from behavioral patterns. An algorithm might learn, for instance, that customers who make more than two visits to a store’s website within a two-week period are 30 percent more likely to make a purchase. Such indicators can trigger tailored offers to convert browsers into buyers, allowing marketers to direct their acquisition efforts and spend toward the most profitable segments.
But buyer behavior has changed significantly since the pandemic began, rendering the relationship rules baked into many existing data models invalid. Externalities that once seemed incidental, such as customer mobility, now have outsize importance. Is visitation down because customers can’t get to the store or because they no longer wish to shop there? Many marketing teams simply don’t know. A Fortune 100 CMO said, “The indicators for the new opportunities we face are not contained in our own data.”
In addition, while patterns exist, they are harder to discern—and even when discerned, they can feel ephemeral, such as communities opening up only to lock down again. To tease out salient behavioral indicators in time to act on them, marketers need continually refreshed data from a variety of sources and at a far more detailed level—looking as deeply as the city-block level in some cases. However, many companies tend to rely on internally derived customer data, using modeling tools that were not built to handle large volumes of data.
Two other issues compound the challenges facing marketers. McKinsey data show that marketing budgets have been slashed for most companies, with six of ten marketers reporting major cuts. “My budget has evaporated,” said one senior marketer. “We have barely enough to execute our ‘must do’ marketing, let alone experiment with new tactics.”
The other issue is the rapid, large-scale shift to remote working. Data-driven marketing works best in agile settings, where teams can test and iterate in sprints. But with nearly two-thirds of employees working from home, marketing leaders have found it difficult to create an effective cadence. “In the past we used to go all in on marketing opportunities by having a command-center-like war room,” said one Fortune 100 CMO, “but with everyone working remotely, we haven’t been able to react as fast as we have in the past.”
How to make modeling more precise when everything else is in flux
While other organizations may have retreated to mass marketing, those that upgrade their modeling can be far more effective in generating revenue. Here’s what they need to do.
Tap new (and better) data
Precision marketing is only as good as the data behind it. New models with old data are still likely to provide inaccurate results. To hone their insights, leaders in the new normal will take a wide-angle approach to data collection by gathering not only behavioral trends and location-based insights but also third-party analytics on their business, customers, and competitors to complement their in-house customer data. Companies starting this journey are finding the most value in incorporating epidemiological data from government sources and customer-mobility and sales data from third-party providers into their models. Companies that extend their data gathering in these ways can identify upticks in demand and where new customers are coming from, as well as assess which customers in their existing base have increased spending and where lapsed customers have gone.
Before it updated its modeling approach, for example, a retail chain could only tell how many customers it was gaining or losing. The company then decided to pull in cell-phone data to scan changes in their competitors’ net traffic. That analysis showed that many of the customers they were gaining during the pandemic were coming from more expensive, specialty players, while those they were losing were heading to cheaper, larger-format players. On the basis of this information, the retailer transformed its onboard and churn-prevention campaigns. They sent emails advertising higher-end offerings to customers transitioning from specialty stores while touting bargain-oriented products to value-oriented customers at risk of churn.
In another example, a business-services provider tapped into new third-party data sources that identify key moments in the small-business life cycle. In one such effort, the provider aggregated data sources that indicated, with a lag of only one day, when new companies were being launched during the turbulence of COVID-19. Their salespeople reached out immediately with products and messages tailored to the needs of newly formed companies, such as systems tools. These collective efforts increased sales productivity by more than 25 percent.
Robust data can also allow companies to generate better competitor insights. By comparing third-party assortment, sales, and promotional data to their own figures, for instance, marketers can evaluate the strength of different value propositions and see which elements resonate with different groups of customers. They can then provide these groups with tailored messaging, content, and offers.
Invest in tech that learns at scale
The increased uncertainty in the new normal requires marketers to get better at testing and faster at reacting. A more agile operating model is a key element in this, but it is also increasingly necessary to work with technology that learns at scale. This requires developing technology capabilities that can read and interpret signals of consumer intent and consumer responses to marketing messages and then feed them back into the marketing engine so it can learn what works and what doesn’t.
Marketers who really push the limits are using artificial intelligence (AI) to monitor campaigns and interrogate responses at a detailed level, to learn not only what works and what doesn’t but for which segments, at what times, and over which channels—and then to adjust their strategy based on those insights. Deriving those specific insights using standard analytics might take the average marketing organization several days. But AI-enabled monitoring can do this in minutes, sometimes seconds.
For example, a consumer services company launched consumer-retention campaigns as communities came out of lockdown. Their customary analytics, which could only assess campaigns in the aggregate, was only marginally effective. However, the organization piloted a new AI engine that could look deeply enough to evaluate responses at the core base statistical area (CBSA), which showed that the campaign was highly effective in specific niches with similar economic and epidemiological profiles. This AI engine will identify how the campaign’s performance patterns evolve, allowing marketers to configure the system so that nightly AI-driven analytics feed directly into the campaign’s targeting logic. This and similar campaigns are a crucial element in a broader data-driven marketing program that has helped the company increase its rate of testing more than fivefold.
Two keys to success: Investing savings and being agile
In order to derive value from these upgraded models, two actions are crucial.
Generate savings to invest in tech
While some companies are simply cutting budgets and retrenching across the board, others are finding it can be more beneficial to reduce spend in unproductive areas and reallocate the savings—as much as 10 to 20 percent of the overall budget, in some cases—into analytics. This requires a thorough but fast reevaluation of all marketing spend to see how the COVID-19 environment has affected ROI. Event sponsorships, traditional TV advertising, and programmatic display based on outdated terms are just a few areas where marketing performance is likely to have shifted significantly. One apparel retailer, for instance, found that the effectiveness of paid search has diminished sharply during the crisis, while social-media activity has been far more productive. Marketing leaders can free additional investment by also reusing and repurposing existing assets. The savings can then be redeployed to fund data-driven growth programs.
Deploy agile marketing in a remote setting
Agile practices are effective in allowing marketing teams to test consumer behaviors and react quickly to changes. While traditionally, agile teams were thought to perform best when working in the same place, the exigencies of the pandemic have required this approach to be rewired for remote work. Leading companies are converting physical war rooms into virtual ones, creating additional points of contact to support adherence to agile protocols (such as sprint check-ins by video, for example) and the use of collaboration tools. The best companies have gone a step further by integrating some of their vendor teams into their remote practices, including working with IT to create shared tools and compatibility guidelines to account for vendors’ different technologies.
Companies that get it right are showing impressive results. For example, a North American telco created a virtual war room that featured online work spaces and digital scrum boards for task and performance management. Sprint planning was conducted using rounds of progressive voting, with each vote spurring active debates on the merits of each test idea. They also held ceremonies using video stand-ups. The virtual war room not only improved test results but helped the team launch tests more than three times faster than their traditional, in-person setup.
Budgeting and operating practices need to be continually reviewed to support this remote agile model. Instead of quarterly or semi-annual planning sessions, marketing leaders should assess performance monthly to ensure that funding and resources are aligned with the biggest opportunities.
Organizations that prioritize their precision-marketing efforts can turn the COVID-19 crisis into a time of transformation. By capturing new data, searching for new behavioral relationships, and enabling rapid experimentation, marketers can seize granular growth opportunities and enter the recovery with significantly greater ROI and resilience.