Market Segmentation means splitting your customer base into distinct communities based on the similarity of their features. Depending on the data you use to segment customers, clustering a market dataset results in the grouping of customers based on geographic, demographic, behavioural and psychographic factors as well as their buying preferences. Powerful techniques like RFM analysis and clustering models can highlight your best customers, direct specific marketing action or prevent churn - and is essential for any data-driven enterprise.
What is Market Segmentation?
Trusting your gut isn't enough anymore. When it comes to customers - businesses are operating in a market of niches where customers have specific interests and mountains of options to choose from. Market segmentation is the process of grouping customers using data on their income, gender, demographics, geography and online behaviour as well as any other valuable features that can help you understand specific communities.
"The mass market is dead, replaced by the mass of niches."
The starting point for effective market segmentation is always - and will remain - customer data. It is part of the reason that data is now considered to be more valuable than oil and why technology companies like Google and Amazon are worth more than national economies.
4 Types of Market Segmentation
Market segmentation can be as simple or as complex as you need it to be. At Graphext, we perform really simple customer segmentation to distinguish between our paid and freemium users but we also have more complex ways of associating customers based on their academic achievements and the roles they perform in their jobs.
Running the risk of stating the obvious - the type of market segmentation that is a good fit for you or your business depends on the kind of insights you want to find and the type of data that you are working with. For instance, you will need data on where your customer base lives in order to find geographical patterns.
"Without a good question, a good answer has no place to go."
But broadly speaking, we've found that these 4 areas of market segmentation are a good place to start and can expose some pretty useful perspectives on your customer base.
4 Types of Market Segmentation; Buying Preferences, Behavioural Segmentation, Geographic Segmentation & Demographic Segmentation.
Your customers put their money where their mouth is ... or rather - what they decide to spend their money on is usually a great indicator of their interests, preferences or attitudes. Segmenting a market based on the buying habits of customers means grouping customers based on what products or services they are buying and what kind of purchases they frequently make.
The kind of communities we might expect to expose by associating customers according to their buying preferences include best customer communities or customer personas that respond really well to email marketing campaigns. What's more, we can also use market segmentation to understand the opportune times that specific customer demographics are likely to purchase items.
Start grouping your customers according to the products they buy and how much they spend.
Behavioural segmentation is similar to the above - in that the behaviour of customers often concerns the buying actions they complete. Nonetheless, effective behavioural segmentation will often require more detailed data points regarding the activity of a customer on an e-commerce site - the kind of event microdata that can be retrieved with Google Tag Manager and seen in tools like Mixpanel or Google Analytics.
This type of market segmentation can help businesses to make predictions about the activity of their customers. For instance, using behavioural segmentation we can build up a picture of the customer persona that is more likely to leave a product basket full of items for longer than a week or, in contrast, the type of customer profile that is more likely to buy an item immediately.
Understanding the unique behaviour of your customer base is one of the best ways to ensure your marketing strategies are specific and relevant.
Internet sales connect distinct markets from around the world. It is now possible - and advisable - for a fabric manufacturer in Asia to cash in on their products via consumers in the USA. But just because businesses can sell to global markets doesn't mean that they operate in a monolithic market.
The kind of customers we find in different geographic markets are as distinct as their cultures, languages and cuisines. Segmenting a market based on location data means discovering the specific features or characteristics of customers in different locations. It really is as simple as that - more detailed geographic segmentation would involve a closer inspection of smaller geographic areas like districts of a town or even PO Box segmentation.
Remember that geographic segmentation will require some kind of variables pinpointing the location of your customers. These might be addresses or cities but most useful and transferrable is coordinates specifying the latitude and longitude of a customer.
Customers living locally to one another can often share similar traits or characteristics. Geographic segmentation means grouping them using geographic data in order to find patterns about location-specific customer personas.
Something most businesses are already familiar with - data-driven or not - is grouping customers according to their age, sex and background. Pivot tables were created with the idea of filtering and retrieving descriptive statistics about specific data segments. This is a simple form of segmentation that can be helpful to practically any marketing campaign under the sun.
But demographic segmentation can be utilized far more effectively than this depending on the detail of data you are collecting on your customer base. Metrics on factors like income, ethnicity, number of children, job type and academic background will help to generate more specific clusters of customers that you can use to direct specific marketing action.
But as much as you want to grow your data-driven market segmentation strategies - don't forget that your customers are protective of their data. Data privacy laws may affect the number of demographic data points that you can collect on your customers.
Demographic segmentation can be as simple as understanding the breakdown of men vs women in your customer base. But bear in mind that the more specific the data you are working with - the more specific the insights you'll find.
Techniques for Market Segmentation
Which techniques you use to segment customers in your market will largely depend on two factors.
- The complexity (or detail) of the insights you need.
- The type of customer data you have to work with.
An example of advanced market segmentation might involve joining data from multiple sources in a knowledge graph to understand the relationships between anger expressed in customer reviews and the time that products were purchased.
An example of simple market segmentation might involve finding the average money spent by female customers. Both methods provide insights - so both are valuable avenues to explore. The difference? One requires much more heavy lifting than the other.
Data types and goals determine how to choose a market segmentation technique.
Here's 3 of our favorite - and most useful - ways of performing a market segmentation ...
Clustering models will group your customers according to the similarity of their features or characteristics. This is a really useful - and powerful - machine learning technique that retailers use to find patterns amongst different customer profiles and to associate customers with one another.
Clustering customers is generally a flexible method of creating market segments and most technologies will allow you to set the customer features used to calculate associations. This helps to place your analysis in different contexts. For instance, starting out with an aim to conduct geographic market segmentation and a question like "Which city do my best customers live in?" - would likely involve associating your customers with values relating to their purchase history and location data.
Network visualisations that link associated customers are a very useful method of exploring the patterns behind different customer profiles, personalities, demographics or communities.
Read more about the different methods of clustering customers in this article.
What customers say matters. Analyzing the content of customer reviews is a hugely important part of market segmentation in the realm of online sales. Pure text analysis of review content is a form of NLP (Natural Language Processing) and wouldn't strictly be considered as customer segmentation.
Nonetheless, associating reviews with specific customer profiles provides a wealth of additional value to your analysis and allows you to start understanding which types of customer personas are likely to react with joy when you offer them a discount code.
It's possible to conduct a simple review analysis without using machine learning NLP algorithms - but we wouldn't recommend this approach. There is so much value to be gained from associating customers based on the sentiment of their reviews or classifying market segments that often react angrily in product reviews that we strongly encourage you to spend more time devising advanced ways of analysing text in your customer reviews.
Read more about how to conduct text analysis of customer reviews in this guide.
RFM stands for recency, frequency and money spent. At its core it is a type of analysis intended to distil the most valuable characteristics of customers into three - easy to understand - metrics.
RFM analysis for market segmentation operates in a similar fashion to the way that clustering customers works. Networks of similar customer profiles are created or metrics given to rate customers based on their value for each of the three key metrics - recency, frequency & money spent. Only with RFM analysis, you would expect to use these metrics alone to calculate relationships between similar customer profiles.
We love RFM analysis at Graphext because it gets right to the heart of what drives a business. Identifying the most productive market segments can drive better decision-making and channel marketing efforts into a straightforward and relevant strategy.
For instance, using RFM analysis- businesses might hope to find the type of customer profile that is unlikely to make purchases after spending a year as a member of the company. This kind of insight would help the marketing team stop sending out email campaigns to this community. A lot of time can be saved with effective RFM analysis.
Read more about the usefulness of performing RFM analysis here.
Tools for Market Segmentation
As you might have guessed from the previous paragraph on our favourite techniques for customer segmentation - it can be a pretty complex process. All three of our techniques - clustering customers, RFM analysis and review analysis - generally involve some form of machine learning algorithm and require a data scientist's expertise to execute.
Despite the prevalence of data science for business and marketing strategies in today's economy - this can often be a deal-breaker ... we know!
Tools for Market Segmentation; BI Tools - Python & R - No-code Data Science
Market Segmentation using Programming Languages like Python or R
If in doubt - get the experts out. It will cost you time, resources and effort but it can be extremely useful to call upon the expertise of data scientists to conduct your market segmentation. We'd recommend this option for larger companies that already have some insights and want to investigate the customer personas in specific communities with precise detail. A data team will be able to handle more complex or creative analysis types than any regular BI tool might be able to.
Market Segmentation using BI Tools
BI tools are great because they offer a way of business analysts with domain expertise to get their hands dirty with customer data. Unfortunately, tools like Power BI or Tableau cannot conduct machine learning analysis or associate your customers with one another in any meaningful way. We'd recommend this solution for any size of company looking to perform basic market segmentation, inspect descriptive statistics about their customer KPIs and explore general market segments with simple charts.
Market Segmentation with No-Code Data Science Tools
In recent years, no-code tools have made significant advancements in computationally heavy industries like web development and data warehousing ... so why not data science. These platforms empower domain experts to execute a strategy from start to finish.
No-code data science platforms let marketing professionals and business analysts conduct complex analytics or market segmentation without having to write code or spent days moving back and forth with a data team. For instance, with Graphext, clustering or predictive models can be deployed on customer data by answering a few simple questions. Then, network visualisations to explore the key features of specific customer communities are created automatically. They won't solve every use case for you ... but can save you a lot of time.
Examples of Market Segmentation in Action
Here's a hand-picked selection of market segmentation analysis that we feel demonstrate the most crucial techniques for finding patterns amongst communities of associated customers.
Supermarket Segmentation: Clustering 1000 Supermarket Customers
Our team segmented 1000 supermarket customers in order to pick out key customer personas from data based on their income, demographics and geography. We isolated the most productive market segment with a clustering model and found that customers in this community were mostly from one geographic region and tended to frequently purchase expensive food and beverage items.
Graph & Sidebar Charts: Customers in cluster 10 have the highest overall spending.
Market Segmentation with RFM Analysis
Menhir clustered a portfolio of 8229 financial assets using an RFM analysis. They used the analysis to identify the most productive community of assets and to report their insights to clients. For instance, pursuing avenues for demographic segmentation, the team at Menhir were able to determine that younger people tended to invest less money than older people.
Analyzing Disneyland Customer Reviews
In this guide, our team analyzed a dataset of 42,656 reviews about 3 Disneyland branches. We deployed out of the box NLP algorithms in order to analyze the topic of customer feedback and to segment reviews according to the business region being discussed. This type of market segmentation let us segment customers based on both their geography and feelings expressed in the customer review.
Conclusions & Further Reading
In short, market segmentation is a powerful technique allowing business teams to associate customers with one another in different contexts. There is little room for doubt that effective market segmentation will improve decision making, help to channel marketing efforts and improve the way businesses identify specific target markets.
Whether you choose to opt for geographic, behavioural, demographic market segmentation or purely associating customers based on their buyer personas - do your homework and make sure to check out the range of purpose built tools that are on offer to help with the challenge!
Market segmentation is a big part of what we do here at Graphext so feel free to look over some of the other investigations we've done into finding patterns amongst customer segments.