“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is written by Marc Vermut, vice president of marketing solutions at Neustar.
In the last few weeks, some marketers have asked me how they should alter their plans to deal with the new COVID-19 reality. Should their advertising go dark? Should they go after new customers or just focus on the most loyal ones? And if they do adjust budgets, how do they figure out where to shift spending?
In this time of great economic uncertainty, none of us knows what’s around the corner. Marketing-mix models haven’t yet measured this world, and the behaviors we’re accustomed to seeing from customers are probably not what’s happening in today’s pandemic environment.
A month ago I wouldn’t have guessed I’d avoid going to the grocery store while changing to a different delivery service. Today, as my once-reliable service experiences major delays, I’ve shifted to a new one.
And it’s not just me: The Wall Street Journal reported on March 23 that in recent weeks, the number of Instacart orders more than doubled, and the size of the orders grew by 15% over the previous year.
This is why marketers need to reevaluate customer propensity. It can provide a sense of clarity and control in a world that so desperately needs both right now. All sorts of marketers, from CPG brands to film studios, retailers and more, are poking and prodding at their customer data to understand the pandemic’s impact on their customers and determine the most efficient ways to refocus their marketing investments.
That grocery delivery service I was using? I’d be willing to bet its data scientists are taking a deeper look at my behaviors to determine the most effective messaging in hopes the brand doesn’t lose me for the long haul.
Like those grocery delivery brands, all marketers should be looking to get a clear picture of audience characteristics and marketing performance when it comes to their new customers, as well as those who have temporarily moved on because their behaviors have been forced to change. If brands are attracting new customers, predictive modeling can help to determine whether they should target those new customers with advertising and what types of messaging speaks to them.
Or, if it turns out brands need to make budget cuts, these approaches can help them determine where to do that in a way that limits negative impact on their business.
Understanding customer propensity
Propensity modeling becomes particularly valuable in uncertain times because it answers a core marketing question: On a scale of zero to 100%, how likely is a consumer to purchase a product or take a particular action? And what will make them more likely to do so? To get to the answer, marketers use prior purchase information such as purchase recency, shopper type and distance from a store.
However, recent retail data suggests that prior purchase information is not necessarily representative of today’s reality. IRI purchase data shows that CPG product category trends are morphing drastically. In the week ending March 15, research showed US shampoo sales rose 42%. Meanwhile, perhaps because we’re staying in more often, hair accessory purchases dropped 13%.
Let’s say you sell baby products. That IRI report showed that in the week ending March 15, baby food and care purchases in the United States were up 63%. Sure, some of those purchases are indicative of panic buying, but the real needs of families with babies and toddlers haven’t changed. It’s their behavior that is changing.
They are, for instance, highly likely to have shifted at least some baby spending online. You can use historical shopping data associated with your identified customers to understand how their behaviors are changing and whether to target them in different ways, and possibly in different channels, as their patterns shift. You want to get ahead of your consumers and be there for them when they arrive.
Movie studios are using propensity modeling to take advantage of changing consumer behaviors – and stay-at-home orders – that prevent people from watching films in theaters. To shift spend efficiently, they’re using modeling to determine who is likely to rent their movies and where they are best reached. It also helps them evaluate the smartest ways to move money originally designated for marketing theatrical releases to marketing home viewing options.
Even smaller brands can gauge propensity with single-channel testing
Perhaps the simplest way to measure propensity is to test marketing in a single channel. Even marketers that do not have a data management platform or sophisticated customer data connecting shopper identities to marketing exposure and purchase behavior can leverage some customer data. They can upload customer information such as emails to publishers such as Facebook and Google to identify audiences and conduct cost-efficient A/B tests.
They might select a subset of diaper buyers and target them on Facebook, then determine how the ads performed to understand their underlying propensity and lift from those ads. This can help determine the value of marketing toward likely audiences or only to those likely audiences that respond incrementally to that marketing.
It’s also where message testing can come in. Does it work better with a given audience segment to emphasize quick delivery, lower price or a loyalty offering?
Even with these single-channel A/B tests, marketers can get a better handle on propensity to buy and measure the incremental impact of marketing. It’s critical for brands to understand which core marketing and advertising tactics they need to keep to maintain and drive incremental revenue while freeing up marketing dollars to invest in new audiences.
We don’t know how long the impact of the pandemic will last. A brand could suffer if it doesn’t have the lights on but its competitors do. During this tough time, deeply understanding and revisiting propensity analysis of customers and prospects can help brands understand changing consumer behavior, retain loyal customers and spot and retain new ones, in the most efficient ways possible.
Follow Neustar (@Neustar) and AdExchanger (@adexchanger) on Twitter.
