There’s lots of discussion among marketers about the best ways to reach consumers in various generational groups. For example, it’s well known that most baby boomers respond better to direct mail while most millennials and Generation Z prefer text and email. Though that’s a great starting point, generational differences don’t always universally apply. The most effective marketing strategy is tailored to the individual. The solution seems simple—but it really takes a sophisticated technique that uses advanced analytics to get these insights. With that in mind, Alliance Data’s Analytics and Insights Institute developed an approach that uses machine learning-based predictive analytics to customize communication channels to reflect individual consumer preferences. After extensive testing and incubation, the team generated a nearly 25% improvement in response rate without increasing marketing spend. This article outlines the strategy and technologies behind this methodology and quantifies the results that can be achieved.
In today’s competitive retail environment, brands need to be effective and impactful at every consumer touchpoint. Large-scale blast communications are a thing of the past, and are quickly proving to be both inefficient and ineffective.
Alliance Data’s recent studies, 2017’s The Generational Perspective and 2018’s The Rules of NextGen Loyalty report, revealed that different generations prefer to engage with brands in different ways. But basing marketing strategies strictly on generational propensities isn’t granular enough. As communication channels continue to evolve and expand, it’s important to be strategic about where to allocate marketing budget to get real return on investment.
Consumers expect you to know them—and how they want to be reached
Marketers often use research-based generational preferences when customizing their messaging strategy. For example, research shows baby boomers tend to respond better to direct mail while millennials and Generation Z largely gravitate toward text and email. Though that’s a great starting point, generational tendencies don’t always universally apply. How can marketers develop a personalized approach that’s scalable and drives revenue?
Alliance Data’s Analytics and Insights Institute has developed a sophisticated technique that applies machine learning-based predictive analytics to tailor communication channels to the individual. After one year of testing and incubation, the team found that nearly a 25% improvement in response rate is possible with less than half the marketing budget, compared to mixed-channel communications based on generation alone.
Tap today’s technology to engage tomorrow’s consumers
As younger generations expect more elevated, personalized experiences across the entire shopping journey, it’s imperative that brands move away from traditional blast-style communication approaches.
The Institute’s two-stage methodology paired historical communication channel behavior, transactional data, demographic data, historical campaign data, and other relevant data sources with advanced data mining, machine-learning techniques, and statistical simulations to arrive at optimal communication strategies that help brands engage with different consumer groups in the channels they prefer most. After more than a year of thorough testing and development, the team found this technique significantly boosted response rate while controlling the messaging delivery investment.
Stage 1: Establish the preferred channel of communication
The first stage used historical transactional data, campaign data, and customer demographics—as well as lifestyle data—to predict each customer’s preferred channel of communication. The most critical component at this stage is to make sure the data is unbiased in terms of the channel of communication, considering how the customers were communicated to in the past. To ensure this, the Institute conducted an experiment that randomized the communication channels and captured consumers’ responses. Once that data was cleaned, machine learning and predictive analytics methodologies were used to assign “preference scores” for each consumer and channel modeled (see Figure 1).
Stage 2: Optimize marketing spend
The second stage used advanced machine learning, simulation, and optimization techniques to achieve the highest response rate at the individual level—all while ensuring marketing funds were fully maximized. For example, while baby boomers may overall prefer direct mail, millennials email, Generation Z text messages, and so on, the reality is that individuals within each of these groups have different individual preferences. The algorithm developed for this stage automatically determined which channel had the most impact for the consumer in question. And that ultimately delivered the highest overall response rate, with the minimum investment necessary, to get results.
An optimized approach drives real results
The approach, which was deployed and tested across several specialty apparel brands, resulted in an investment savings that ranged between 30–40% compared to traditional methods (see Figure 2), and a 15–24% lift in response rate.
Figure 3 illustrates how the channel optimization approach stretches marketing dollars further by improving response rate. Theoretically, higher response rates can be achieved by increasing marketing. In reality, marketing budgets tend to be fixed and real value can be unlocked by utilizing the budgets in a smarter way. By implementing an optimized communication strategy, marketing budget can be allocated to the right message delivery tactic for the right consumers at the right time.
For example, this strategy can differentiate the baby boomers who respond to direct mail from the ones who prefer digital methods. On the other hand, it can also separate and target the smaller percentage of millennials who actually prefer direct mail over digital communications. By implementing an optimized communication approach, marketers can determine where their marketing spend will make the most impact.
Predictive personalization is the future of retail marketing
As brands aim to anticipate consumers’ evolving needs, continuing to optimize marketing strategies will require the ability to deliver deeply personalized messaging, offers, and experiences through advanced data practices that include predictive modeling, machine learning, and artificial intelligence. It’s no longer enough to rely on traditional, blanketed marketing approaches and expect big results. Marketers who are willing to dive deep into the data will have the insights they need to maximize campaign dollars and drive real growth for their brands.
It takes time, capital, talent, and a robust testing plan to garner the deep insights needed to drive real return on investment. A strong marketing strategy is built on solid predictive analytics that use modeling to reach consumers through the channels they use most. The key is to strategically develop these techniques from the outset—or find the right analytics partner to help set the pace for future success.
DATA MATTERS. IT DRIVES EVERYTHING WE DO AT ALLIANCE DATA, AND IT ALWAYS HAS.
The Alliance Data Analytics and Insights Institute is a data-driven, in-house consultancy providing world-class expertise, resources, insights, and thought leadership for our valued brand and industry partners. Our proprietary and industry-leading toolsets, combined with our experienced data scientists, strategists, and thought leaders, enable us to reveal data-driven, actionable insights that help brands understand their customers better, drive increased loyalty and engagement, and ultimately grow their businesses.