Customer segmentation models

GrowthLoop Icon
Researched by
GrowthLoop Editorial Team
verified by
David Joosten

Key Takeaways:

  • Customer segmentation helps marketing teams create personalized, targeted marketing.
  • There are many different kinds of customer segmentation models that work for different industries and companies.
  • Experimenting with multiple customer segmentation models helps teams optimize campaigns and ROI.
  • A customer segmentation model may need to change over time.

Table of Contents

What is customer segmentation?

Customer segmentation is the process of dividing a large group of customers into different groups, based on characteristics the people in those segments share. Different people may be interested or attracted to different things, which is why marketers often group customers into “segments” and tailor marketing strategies to those groups

These characteristics may include location or age group data, values and preferences, needs, or technological capabilities. Segmenting customers makes it more effective for a marketing team to reach specific groups of customers or potential customers with personalized messaging, upsell or cross-sell offers, and improve customer retention and loyalty.

For example, a pet food brand has different needs-based customer segments for customers with dogs and customers with cats, so they can send only their dog-owning customers messages about their dog products.

What is a customer segmentation model?

A customer segmentation model describes how customers get subdivided as part of a marketing strategy. For example, a demographic segmentation model is a traditional model that divides people into groups based on their ages, genders, sexual identities, and ethnic or racial backgrounds. This means you can tailor advertising to groups like, “white women aged 50 and older” or “unmarried men between 25 and 35”.

The model you use depends on your business, the types of data you have access to, and how people interact with your products or services. The demographic model can be useful for a brand aimed at those specific groups, but it’s less useful for one that caters to everyone regardless of their background, like a gas station or grocery store. An organization in that situation may pick another model, such as a geographic or needs-based model.

What type of data is used in customer segmentation models?

A customer segmentation model will take in a variety of data points. These data points will depend on the model type. For example:

  • If you were building a technographic model for your website’s users, you need to collect the types of devices they use: Mobile or desktop, Android or iPhone, Windows or Mac.
  • Demographic or geographic data can come from website registration or other customer-facing forms on your website.
  • For a model based on behavioral data, you could track clicks or cookies on your website, but if you wanted to capture psychographic data you may need customers to fill in surveys or answer questionnaires.

Customer segmentation models benefit from having a single source of truth for customer data, which is typically stored in a cloud data warehouse.

Types of customer segmentation models

There are many ways to divide a group of people into smaller segments, and some of them will work for your brand better than other models.

The following are the most common types of customer segmentation models:

Demographic

Demographic information is population-related data, such as age, level of education, gender and sexual identity, racial or ethnic background, income, relationship or parental statuses, and similar characteristics. Demographic models create groups based on broad similarities between people in those categories, such as men or women of a certain age, or “luxury” brands aimed at folks in high income brackets.

Behavioral

Behavioral information includes past observed behaviors that may indicate a customer’s decision-making process or possible future interactions with your brand. This data may include their purchase history, social media platforms of choice, and other habits. For example, behavioral segmentation may include customers who have purchased your products or interacted with your brand’s social media.

Psychographic

A psychographic model considers a customer’s beliefs, attitudes, habits, personalities, or interests. If a customer survey asked, “Do you agree with the following statement?” and then presented a belief or idea, the customer’s response to that question may put them in one segment or another. This may be a useful model for lifestyle brands trying to attract customers who align with the brand’s philosophy.

Geographic

The geographic model divides people based on where they are located. This may include data like regions of a country, like the East Coast of the United States, or smaller regions, like towns or zip codes. It may also consider whether your customer is located in a suburban, rural, or urban area. Climate differences could also be a variable, such as areas that get snow or have dry seasons.

Technographic

Technographic segmentation depends on the devices and technological services your customers use. Data for this model may include the customer’s variety of mobile devices, social media services, and software or hardware preferences. For example, video game developers may create a technographic segment that includes only people using the consoles where their games are available.

Firmographic

Many business-to-business (B2B) services will use firmographic segmentation to divide the companies they want to work with into smaller groups. Firmographic data includes information like the company’s size, industry, location, growth trends, and ownership structure. For example, a software company targeting small businesses may target smaller size companies than a software company targeting enterprise public corporations, and firmographic data can help with that differentiation.

Needs-based

Needs-based segmentation models target customers who require a specific type of product or service. Returning to the pet food brand example, their needs-based model would send dog-related product offers to dog owners, but not customers who don’t own dogs.

Value-based

Value-based models focus on successful sales and existing customers. These models will look at the customers who have had the most repeat purchases or the highest returns on investment (ROI). Placing this focus on existing customers can help your organization target marketing to people who already support your brand, which helps build trust and loyalty. A Harvard Business Review study found that it costs five to seven times more to acquire a new customer than to retain an old one.

Highly targeted customer segmentation models

Depending on the product or service, industry, and available data, the customer segmentation models previously listed can sometimes be too general for highly personalized marketing. That’s where more specific segmentation models can be helpful. Some examples of these models may include:

  • Customer lifecycle: The customer lifecycle is a way to map how a customer journeys from considering a product to becoming a customer. A customer segmentation model based on this map would include segments for customers at each stage of that process. One segment may include customers who have never heard of your brand (awareness stage), and another may include repeat customers (loyalty stage).
  • Recency-frequency-monetary value (RFM) models: For an RFM model, consider the recency (how recent was the customer’s last purchase?), the frequency (how often do they make purchases?), and the monetary value (how much do they spend on each of those purchases?), and then use those groupings to identify high-value customers (HVCs).
  • Cohort-based: “Cohort” is a general term for any group being observed over time. This means you can define a metric within your own data, and create customer segments based on the cohort of customers who fit within it. This could include data points relevant to other segmentation models, or data you have that doesn’t fit within another model. For example, a cohort could be defined as customers who made their first purchase in December of 2023. Conducting a cohort analysis allows you to identify trends in that cohort over time (such as “what happened to these customers?”), which can help refine your customer segmentation model.
  • K-means clustering: This type of clustering is a statistical algorithm that identifies data points with similar characteristics, and places them in clusters together based on how close those characteristics are to each other. This is an algorithmic model that may create groups outside of the standard segmentation models mentioned above.

Customer segmentation model examples

The value and efficacy of a customer segmentation model varies across organizations and industries. For example, a fintech B2B company will rely more heavily on models that focus on companies as customers, which could include a firmographic model or a geographic model. These models could include:

Firmographic

Annual Revenue:

  • $10,000
  • $100,000 
  • $1 Million
  • $1 Billion

Company Size and Age:

  • “Small”: 1-3 years in operation, 1 - 100 employees
  • “Medium”: 4 - 10 years in operation, 101 - 1,000 employees
  • “Large”: 11 + years in operation, 1,001+ employees

Geographic

US regions:

  • “Northeast”
  • “Southeast”
  • “Southwest”
  • “Midwest”
  • “Rocky Mountains”
  • “Pacific”

If a fintech company is trying to market a $15,000 fraud protection service, these segments provide some useful options:

  • A firmographic model - Companies in the million- or billion-dollar segments are more likely to have a budget for the services; the $10,000 group may be priced out.
  • A geographic model - If New York State passes regulations requiring businesses with offices in New York to retain a fraud protection service, the fintech company can send marketing to the “Northeast” region about this regulation.
  • A firmographic and geographic model - If this regulatory requirement has an exemption for companies with fewer than 100 people on staff, the fintech company can further narrow its focus to exclude the small size companies..

A business-to-customer (B2C) company may also find other models more descriptive of their customers, such as a demographic, psychographic, or technographic.

Demographic

Age and gender:

  • Males aged 25 - 45
  • Females aged 40+
  • Males and females aged 15 - 20
  • Any gender, 21 - 29

Psychographic

Activities, interests, and opinions (AIO):

  • Running or jogging
  • Team sports
  • Health and exercise
  • Hiking and climbing
  • Fashion

Technographic

Devices owned:

  • Desktop computer
  • Laptop computer
  • Android phone
  • Android tablet
  • iPhone
  • iPad or Apple Tablet
  • Other mobile device

For example, an ecommerce company is trying to market a fitness shop app, selling athletic equipment and apparel. So, they could use any of these models: 

  • A demographic model -To attract high school customers, they could aim marketing at individuals ages 15-20, or create ads aimed at the students’ mothers in the “Females aged 40+” segment.
  • A psychographic model - They could send a survey asking what kinds of hobbies their customers are interested in, and send different offer emails to respondents in the “Team sports” and “Fashion” segments.
  • A technographic model - If they want to attract more customers to use their mobile app, they may send an email to anyone who has accessed their website on an Android or iPhone device in the past.

Benefits of using customer segmentation models

Customer segmentation models provide a framework for a customer audience strategy. A good customer segmentation framework allows you to tie audiences to an organized structure, minimize gaps, and helps you address customers where they are. A customer segmentation model has specific benefits for marketing teams, including:

Personalized messaging

Segmentation gives you a clearer picture of your audience, which helps you craft branded messaging tailored to specific customer segments. Instead of trying to create one message that works for everyone, you can create a few messages that are more targeted to different groups.

Upselling and cross-selling

By identifying which segments are most likely to purchase additional products or services, you can focus marketing efforts on those customers instead of toward likely one-time customers. For example, a sports franchise trying to drive ticket revenue could promote a multi-game ticket package to customers who have purchased tickets one to three times in the past, or season tickets to customers who purchased tickets to more than four games last season, and upgraded seats or a suite to current season ticket holders.

Customer retention and churn prevention

By finding the right indicators, you could identify customers at risk for churn, and craft targeted marketing strategies to engage customers and improve their experiences. Your customer success teams could also use this information to engage with those customers sooner and more directly.

Improved customer satisfaction and loyalty

Understanding which channels your customers use means you can target various segments for additional customer service or specific messaging. Understanding your customers better can also help you understand their needs, pain points, and churn risks. This can help you craft a more robust and targeted customer service strategy with your customer service teams.

Personalized messaging, upselling and cross-selling opportunities, addressing churn, and improving customer service are all steps you can take that will improve customer satisfaction and foster more loyal customers.

How to build a customer segmentation strategy

Creating a customer segmentation model begins with developing a strategy for that model. This will help you narrow down what type of model you need, and what data you still need to collect before implementing that model.

Set a goal

Begin by setting your marketing goals and determining which kinds of customers you want to target with your customer segmentation strategy. Identifying possible key performance indicators (KPIs) at this stage is useful, but it may be easier to do this after you have developed a model.

  • Are you trying to land higher value customers, or develop more repeat customers?
  • Are you trying to fulfill compliance or regulatory needs, or are you trying to position your brand as a must-have luxury?
  • What is your current largest group of customers?
  • What assumptions are you making about your customers? Are those assumptions biasing your campaigns?

Create a single source of truth

You need a single source of truth for your customer data; one place where all of your data is collected and collated, such as a cloud data warehouse. Once you have that, look at your data to determine which segmentation models you can create. What kinds of data are you already collecting about your current customers? What kinds of data would you need to meet your goals and ideal segments, and how can you collect that data?

Identify key attributes

The next step is to identify key attributes for defining the different segments. Identify two or three key attributes, and use them to divide the customers into different categories. For example, geographical region or age. Select criteria that are most useful for your business service or product.

Begin with broad criteria for these categories, and then narrow the parameters. For example, if you segment customers by age into “over 40” and “under 40” gives you two groups. But you may find it more useful to have “under 30”, “30 to 50” and “over 50”, or even more granular groups than that.

You want categories that are mutually exclusive and comprehensively exhaustive (MECE). “Mutually exclusive” here means that nobody in one category should appear in another category, and “comprehensively exhaustive” means that the options should cover all possibilities, and nobody should be left out of any of these categories. 

While it is possible to create more granular segmentation with more than three key attributes, adding more criteria also adds extra levels of complexity that can make the segmentation model more difficult to use than the potential benefits.

Apply the model to your data

The final step is to apply your customer segmentation models to your data, and discover how those models subdivide your customer base. It may be helpful to create a test case by taking a small group of customers, predicting which segments they’ll appear in, and then verifying that the segmentation model worked properly.

How to drive adoption of your customer segmentation model

Having a customer segmentation model is helpful, but it’s only effective if teams use the model. It can be challenging to have the organization adopt the segmentation model, but these tips can help you drive adoption:

  • Executive support - Getting your executive team on board helps garner top-down support for the model, so prepare a presentation for them to explain the model, your strategy, and any relevant data you discovered while creating these models.
  • Test audiences - Use the model to create a test audience, and then show that new audience to your marketing teams. If you can provide them with a larger, more receptive, or otherwise more valuable audience than their current strategies have decided, this may lend some bottom-up support for your models.
  • Identify concrete improvements - When talking about the model with other teams, it helps to give them a specific and relatable example of why the model is better. This may require asking those teams what kinds of KPIs or customer insights they rely on, and then finding ways your new segments apply to that data.

How to know if you have an effective customer segmentation model

It may take a few attempts to land on a customer segmentation model (or models) that works for your marketing team. Your KPIs are a good start, but will only tell you part of the story. When determining whether your model is working, keep the following questions in mind:

  • Is the model comprehensive? Does it apply to all of your current customers? Who is it leaving out?
  • Is the model simple? Do your marketers understand the model? Is it difficult to explain to your teams?
  • Is the model effective? Can you create personalized marketing campaigns with it? Can your teams create 1:1 experiences with it?

Over time, you may find that your customer segments don’t give you the results you want or that customer needs have changed. Check your segmentation model against updated goals. If the segments aren’t supporting those goals, it may be time to reevaluate them.

Analyzing customer segmentation models

People and needs change over time, which will impact your customer segmentation models. So, checking your models periodically is critical to ensuring they continue to work for you. The cadence for these check-ins may vary across industries; some customer needs won’t change much over a few years, while others’ may change within a few months.

Your customer segmentation analysis should include the following steps:

  • Perform an internal needs assessment to determine if the models still serve your strategy.
  • Prioritize and measure your KPIs for the model against your goals.
  • Gather and review customer feedback about marketing targeted for their segment.

Tools for customer segmentation

Creating a customer segmentation model can be a manual process, but some tools on the market can automate parts of the process. Generally, a customer segmentation tool taps into your customer data, so you can use it more effectively to create market segments. Some also include tools for creating audience lists based on your identified segments.

Some of these tools include:

  • GrowthLoop
  • Twilio Segment
  • Mailchimp
  • Hubspot
  • Sprout Social
  • Qualtrics
  • Bottom Line
  • Experian
Published On:
February 1, 2024
Updated On:
February 1, 2024
Read Time:
5 min
Want to learn more?
Contact Sales
NExt Article
Composable CDP
Previous Article
Marketing automation

Looking for guidance on your Data Warehouse?

Supercharge your favorite marketing and sales tools with intelligent customer audiences built in BigQuery, Snowflake, or Redshift.

Get Demo