How to use choice modeling surveys to predict customer behavior

What happens when you don’t know exactly what your customers want? Thanks to rapidly developing technologies, increasingly competitive markets and more products and services to choose from, understanding customers’ preferences and predicting behaviors is harder than ever before. That’s where choice modeling comes in. As an established methodology to estimate the decision making processes and preferences of different customer segments, it will be invaluable for you if you’re just starting up, making changes to existing products and services, launching new products and more. Read on to learn more about what choice modeling is and the tools you can use to do it.

Let’s dig a little deeper into what choice modeling is and how you can make the most out of it. 

Essentially choice modeling is an experimental method used to understand how individuals and groups of individuals or customer segments make decisions among discrete alternatives. Imagine that a customer is inside a grocery store, looking at shelves and shelves of different toothpaste. How do customers choose which toothpaste they want?  The variance in toothpaste features alone is vast—there are whitening toothpastes, toothpastes for sensitive teeth, different flavors, different sizes, different levels of fluoride and more. In this case, choice modeling allows you to explore customer preferences of numerous different possible configurations and design alternatives in order to understand the optimal combination of features that matter most to consumers. 

For example, you might present the following three alternatives to a sample of 100 toothpaste consumers:

These three combinations are known as the choice set

Choice modeling essentially predicts consumer preferences for the different buckets of features from the choice set. This is a simple example, but in fact choice modeling is even more powerful when it comes to more complex models where the combination of features is numerous. For instance, it is possible to generate choice sets that combine the features of the toothpastes with the features of the shoppers, such as their income level and age. Choice modeling, then, helps businesses to understand the factors that shape decisions, and allows them to estimate demand for products.

Answers to questions like these are invaluable for planning, marketing and budgeting purposes. These answers can also help you to better communicate with your customers, and increase the odds that your marketing efforts will pay off.

Long-term business success is not just about offering a great product or service to your customers—but that’s definitely important. It’s about understanding what makes customers tick, meeting their needs and engaging with them in ways that resonate. In order to do this, you’ll need to understand customer behavior such as the habits, preferences and psychology that drives customers to make buying decisions.

Here are a few different types of consumer behaviors:

Clearly, this is powerful data. If you know what your customers’ needs and preferences are, how they shop and consume your products and the factors that drive their buying behavior, you’ll be in a great position to create a positive customer experience, foster a loyal customer base and to beat off the competition. 

So, understanding customer behavior is immensely powerful, but how can you possibly predict it? Let’s take a look at how market research and customer behavior surveys can lay the foundation for gaining critical insight into your customers and their habits.

Market research is crucial for learning more about your customers and their behaviors. Market research essentially involves gathering, analyzing, interpreting and applying a range of information to build a picture of the segments buying your product or service, the size of the market, your customers’ habits and their needs and desires. The types of market research that could potentially be conducted are numerous, but simply speaking, there are two types of research you should be using:

Exploratory market research is non-specific in nature and helps to build up a broad picture of your market and gain insights and opinions on your product, competitors and the consumption process.  Alternatively, your primary research might have a specific focus. For example, if you want to understand exactly how customers move around the aisles in a grocery store, you might conduct some focus groups to ask customers how they would engage with specific store layouts. This sort of information is useful for customer behavior analysis, and in shaping future survey development. 

For example, performing analytics of your own sales platforms is invaluable in telling you things like how much customers typically spend, when they spend and what they're buying. Many other aspects of customer behavior can be captured through analytics of your marketing channels. Evaluating website page views, clickthrough rates and conversion activity can tell you a whole lot about what customers like and don’t like, and what marketing content works. You can also leverage competitive intelligence in order to understand more about the behavior of prospective customers. 

Customer behavior surveys can be employed to learn more about your customers’ habits and  behaviors. These are surveys that are designed to gather demographic data about your customers, their preferences and desires, any needs that are currently being unmet and their purchasing patterns. 

There are many different types of data that you can collect through your customer behavior survey. For example, BridgeCare Finance, a new startup that helps families pay for childcare costs, used consumer behavior surveys of parents of children under age five to learn about the type of debt parents face, the problems that they currently have with childcare costs and crucially—to learn about whether parents would be interested in BridgeCare’s new services. 

In particular, we recommend two types of consumer behavior surveys:

Learn about what they currently buy or have bought in the past. This can help you learn more about what customers like, what features they’re looking for in a product or service, and what drives them to make  purchases. This is vital information when operating in a competitive market.

If you also vary features about the product, such as the price, and repeatedly survey them, you can build up powerful insight into the patterns that prompt immediate purchase and that lead to delayed purchases—crucial marketing knowledge.

The type of consumer behavior survey described above, in which variables are adjusted and customers repeatedly surveyed, is essentially a simplified way of presenting customers with different choices, and then determining the impact on purchasing patterns. That is useful, but what happens when the potential combination of possibilities is seemingly endless? That’s where choice modeling comes in. In the age of Big Data, choice modeling is a powerful way of predicting customer preferences and behaviors, and improves the accuracy of your predictions. 

For marketing activities, choice modeling is especially useful for examining the following:

By presenting customers with a range of different options and estimating the likely demand for each, you essentially have the ability to gauge market share. And, this works not only for market share between your bundle of products and services and those of your competitors, but also the share that each of your own products contributes to your overall sales. This information can be used to build fairly accurate revenue models, which in turn are useful for forecasting sales, and planning marketing activities.

Let’s imagine that you perform a choice modeling experiment that asks customers to indicate their preferences among six different scented hand lotions. Your research indicates that 60% of customers prefer the lavender fragrance, but only 5% would choose a sandalwood fragrance. Information like this is hugely useful in informing your promotional campaigns. For instance, you might offer a buy one get one free offer for customers buying sandalwood in order to push sales, or invest most heavily in the placement of the lavender hand lotion.

Choice modeling helps you compare customer preferences for different bundles of product features, and that can be used to support positioning strategies. In the example above, the known unpopularity of sandalwood can be used to support reconfiguration of the product in order to boost its appeal. Alternatively, combining this data with data about the customers that do like sandalwood can be used to support product positioning—sandalwood might be marketed as a fragrance that appeals to older customers. 

So, by using choice modeling you essentially gain access to some very comprehensive data that will allow you to to predict consumer behavior, develop impactful marketing campaigns and drive sales. Let’s take a look at a specific type of choice modeling and its potential applications.

The two main types of choice modeling are conjoint modeling and discrete choice modeling. Conjoint analysis means that the research participants evaluate the different product configurations independently of one another, which can be used to identify the influence of product attributes on profile evaluations. In contrast, discrete choice modeling asks respondents to consider multiple product profiles at the same time. In other words, participants are presented with a series of different product configurations, and asked which, if any, they’re most likely to purchase.

In this analysis, the influence of the various product features on choices is evaluated. This yields a measure of the relative importance of each feature, as well as a measure of the strength of influence of each feature. This is very powerful, because in that way, you can determine the interaction of different attributes on purchasing decisions. Let's take a look at some applications of discrete choice modeling.

Imagine you are offering your customers different membership options for your gym. There are three aspects of the price. These aspects are the annual, enrollment and monthly fee. Discrete choice modeling can help you to determine whether the attractiveness of the monthly fee depends on the size of the enrollment fee and annual fee, and by how much. This will help you pinpoint the optimal bundle of fees that will appeal to your target customers. 

Imagine sales of a particular product are stagnating. The marketing team suggests adding some new product features, but what features will be especially appealing to customers? Perhaps it's a new color, an updated application, a unique size or something completely different. Perhaps investing in just one of these enhancements will be enough to turn around the fortunes of the lagging product, or a distinct combination will be best. Discrete choice modeling can be used to identify the best strategy for success by presenting discrete bundles of enhancements, and asking them to indicate their preferences. Not only will this provide you with the critical insight you need, but it will also save you wasted investment in product enhancements that your customers don’t really want.

Finally, discrete choice modeling can also help you model the impact of potential competitor moves on customers’ purchasing decisions. For example, you might ask research participants to indicate their preferences between a new product that you are launching and a similar one that might be launched by a rival. If your analysis reveals that your competitors’ product is more attractive than yours, you can make tweaks and adjustments to increase your appeal—even before your competitor makes a move.

Choice modeling can help you understand customer preferences at a deep level, estimate demand and better serve your customers. From product discovery to buying your product, understanding your customer and their buying behavior is imperative. You can further enhance your knowledge by downloading our guide on customers' buying behavior. And, if you’re ready to take the first step towards making the right modeling choice, start by exploring our market research solutions.