Historically, prospect segmentation has been an effective tool to help marketers reduce the cost of retaining customers. But retaining customers is not enough. Your existing customer base is not static, it changes all the time. Customers that stop buying from you must be replaced, and additional new customers must be added to grow your customer base.
The problem with new customer acquisition has always been its cost. But that has changed. Modern advances in prospect segmentation — judicious use of Artificial Intelligence and Machine Learning — have made segmentation an effective tool for acquiring new customers as well.
Now, enterprises and mid-market companies can identify and reach likely new customers with greater cost efficiency than ever before.
Not all your customers are alike, but within the universal set of your customer base there are subsets of customers who share certain characteristics.
Prospect segmentation, also called market segmentation and customer segmentation, is the practice of grouping customers by those shared characteristics.
In the dawn of the industrial age, prospect segmentation did not yet exist, because the means of production was the biggest challenge. Customers bought what manufacturers built. There weren’t many choices.
By the mid-20th century, new production techniques combined with lower costs enabled manufacturers to make diverse lines of products. As soon as customers had choices, it became important to recognize which customers wanted which products.
Availability of data and advances in information technology have enabled marketers to segment a growing body of fields.
While marketers use geodemographic segmentation to identify the locations of likely audiences, they also are used by governments for public planning.
The most rudimentary form of segmentation, geodemographics help identify similar cohorts based on simple data such as:
Knowing what people have bought, how frequently they buy it and how much they’ve spent on it allows marketers to calculate Cost to Acquire the Customer and Customer Lifetime Value.
One technique uses information such as Average Order Volume (AOV), Customer Lifetime Value (CLV) and Customer Acquisition Cost to group customers into segments like these:
Transactional segmentation also identifies when individuals stop purchasing and helps predict when marketing to them will become inefficient. The combined RFM score for each customer makes it easy to measure their relative value against each other.
Another technique gives customers a score in each of three categories:
Buying behaviors can be segmented by any of these four parameters:
A buyer makes a purchase because of an occasion or event. They can be added to a group based on whether the purchase is for a one-time occasion or repeated.
A son in Arizona buys flowers for his mother in Florida twice each year, for her birthday and for Mother’s Day. Each time he buys flowers for her he uses the same florist near where she lives. These are repetitive occasions. He might also send flowers to each of his clients, located all over the country. These are multiple examples of one-time occasions. The florist in Florida would put him in a group of repeat buyers, while the florists serving his business needs would segment him as a one-time buyer.
Whether a buyer is a heavy or light user of a product is important information to a seller of consumable goods, products that require routine service or depreciating assets.
Consider two buyers of the same make and model pickup truck. Buyer A likes the way the truck looks and feels on the road, but uses it primarily as a commuter vehicle and puts 9,000 miles a year on the truck. Buyer B uses the truck as a work vehicle on construction sites every day and racks up 20,000 hard miles every year.
A customer may be loyal to one brand, prefer to buy a different brand each time or start the buying process over from scratch. Whatever segment the buyer fits in, the information is important.
In the Usage example above, Buyer A likes the way the truck looks and feels on the road. But it’s a short-term purchase that will be replaced in three years. And the new truck will always be a different brand. Buyer B has a reliable truck and buys the same brand every time.
If the same dealership sold pickup trucks to both Buyer A and Buyer B, they would know to spend their loyalty marketing dollars on Buyer B, not Buyer A.
Because of the array of benefits available to today’s consumers, they can have a difficult time figuring out which is the “best” deal using objective criteria. By observing a customer’s behavior over time, marketers can learn which benefits are most important to that buyer.
Examples of benefits that can motivate a buyer to act:
The biggest difference between benefit-based segmentation and need-based segmentation is that benefit-based segmentation concerns seller-offered benefits, while need-based segmentation is based on the emotional needs and preferences of the customer.
When you register online with an investment firm, they ask you a series of questions about your preferences to determine how conservative or risk-taking an investor you are. Based on your individual needs, the firm suggests portfolio options that fit your profile.
Here, the buyer is attracted to the firm not so much because of a promised financial return as the firm’s ability to understand and respond to their needs. A large firm will have a complete array of options to accommodate the needs of all investors and the ability to segment their customers by need.
Other examples of needs segmentation might include the need for:
Unlike geodemographic and transactional segmentation, need-based segmentation is highly specific, and segmentation criteria must be customized for every seller. Once those segments have been established, customers can be grouped and marketed to appropriately.
Buyers’ attitudes affect their purchase decisions. Segmenting a customer base by attitudes provides insight into how best to position offers and develop messaging for each segment. The guiding principle of attitudinal segmentation is relevance.
The better aligned your offer is with the customer’s attitude, the more consideration it will receive.
A dental practice’s patients who are most concerned with oral health might receive one offer, while those patients most concerned with dental cosmetics would receive another.
Creating attitudinal segments has long been the work of traditional research companies that use surveys and focus groups to uncover customer attitudes.
All of the above types of segmentation were developed before the existence of artificial intelligence (AI). They can all be performed “manually.” AI represents the next big step forward in segmentation.
The traditional methods of manual segmentation shown above are limited to scratching the surface of data that marketers can access today. Segmentation improves exponentially when data handling is faster and less subject to human bias (which rigs the outcomes to match preconceived notions of what they should be), and when data storage capacity is completely scalable.
Artificial Intelligence (AI) delivers by objectively mimicking human thinking at tremendous speed. Thanks to that speed, objectivity and ability to handle more data, AI is able to recognize more patterns in the data and create more granular and more-precisely-targeted segments.
Here’s a short list of ways AI improves prospect segmentation:
With the addition of new segments and the refinement of all segments, AI tells marketers more about their customers, but the additional data and granular segments add complexity.
AI employs Machine Learning (ML) to handle that complexity and predict which records are most likely to buy.
Machine Learning (ML) and Artificial Intelligence (AI) are often used interchangeably. While the two terms are often used interchangeably, ML is actually a subset of AI.
As with human learning, ML is the product of many experiences. But while human experiences occur over long stretches of time and are informed by emotional responses, Machine Learning algorithms run thousands of data models in very little time to produce objective results. Given the appropriate parameters, ML automatically creates many different models using the segments defined by AI.
The final result is a data model that identifies most likely customers and least likely customers by their segmentation characteristics. To make sure the results are valid, the data is used to train, validate and test the model.
Once the model is ready to use, it can be applied to new records to identify likely buyers who should be added to the list. What’s more, ML can automatically adjust marketing campaigns to deliver the right message to the right customer in terms of headlines, body copy, calls to action, images, colors and delivery time.
Using AI (and ML) in Prospect Segmentation identifies:
This allows marketers to (1) eliminate the least profitable customers from a campaign and (2) add new customers to the campaign database whose characteristics match those most likely to buy. Removing non-buyers from the campaign and replacing them with buyers yields two significant benefits.
The combination of higher response rates and lower acquisition costs means that a marketer can either lower the reach and cost of their campaign while producing the same number of sales, or, budget the same amount on their campaign and produce more sales.
This is why AI is being used more and more for enterprise and mid-market company marketing campaigns.