Customer segmentation best practices: a how-to guide

In this article, we’ll delve into some of the reasons why we’re such big segmentation fans here at Peak, showcasing why it’s so important and how marketing teams can further harness its power to drive significant gains.

Psst — if you’re already pretty clued up when it comes to customer segmentation, you may want to skip ahead a little, but if not, let’s cover some basics first…

What is customer segmentation?

Customer segmentation is the process of creating cohorts of customers into groups of similar commonalities, such as demographics, product preferences and value to a company. Effective customer segmentation — targeting a smaller number of the right customers with the right content — enables brands to increase both their relevance and resonance.

You can’t expect your company’s entire product range to be of interest to every customer every time, but through effective segmentation you can get more of what they’re interested in in front of them more often.

By finding commonalities between your customers, you’re unearthing key connections, mutual interests or shared needs. All of this is imperative when it comes to building brand loyalty and encouraging repeat purchases — in fact, 78% say that personalized content makes them more likely to repurchase, according to a recent McKinsey study.

So, where do you start? First you need to create your segments, and there are a number of ways in which you can segment a customer base. If I wanted to target men aged 25-40 based in Los Angeles, I’d leverage both demographic and geographic segmentation models. Perhaps you’ve got a special discount offer only applicable to purchases made via your company’s app — if so, you’d use technographic segmentation to identify your customers who have your app downloaded to your smartphone.

These are two of the basic methods that will help you to create useful segments of customers — but there’s much, much more you can do to drill down into your customer base to enhance your targeting. Let’s look at a few more useful segmentation models defined by the marketing gurus over at HubSpot. 

Customer segmentation: getting the basics right

1. Demographic segmentation

This is an obvious one, but is more than likely the first port of call when it comes to creating initial segments. You wouldn’t want to waste money targeting a millennial with a product that you know only resonates with Gen Z, in the same way you’d be wasting ad spend pushing women’s clothing to a male audience segment. By factoring in information like age, gender, job title and income you can shape your campaigns to better meet a person’s needs and interests.

2. Geographic segmentation

As Catherine stated in the above quote, always make sure you’re targeting people with information that works in their location. You wouldn’t target a “Visit New York!” ad campaign to people who already live in the Big Apple, for example. Building geo-targeted segments allows you to significantly increase resonance with your customers — you could even try including some regional expressions or local references in marketing to make it feel even more targeted, too.

3. Psychographic segmentation

What matters to your customers? Showing them marketing materials and content that they’re actually interested in a crucial aspect of your marketing — there’s no point running a World Cup-themed campaign if you don’t have any customers who care about soccer, for example. Tap into their interests, values and personality traits to further resonate with who you’re speaking to.

4. Technographic segmentation

Desktop or mobile? Chrome or Firefox? Mac or Windows? These types of technographic information can help you not only gain a clearer understanding of how your customers browse, but also allows you to tailor the experience they get from your brand. Given that emails will often appear differently depending on which email provider they use, for example, this is a crucial step that is often overlooked.

5. Behavioral segmentation

This type of segmentation is useful for things like basket abandonment emails, and for gaining understanding of how your audience is currently feeling about your brand. If they’ve just left a favorable review online, for example, they may be more receptive to some further related product ads to complement their previous purchase. A negative review, however, means you’d likely have to tailor your approach to be a little softer.

6. Needs-based segmentation

A good example of needs-based customer segmentation would be segmentation that factors in specific characteristics that perhaps influence the products they’d require from you. For example, a campaign promoting left-handed golf clubs would be wasted on right-handers, and vice versa!

7. Value-based segmentation

At Peak we place a strong emphasis on value-based segmentation, which we’ll go into further in the following section. But in summary, you want to identify segments of your most valuable customers, in terms of factors like propensity to purchase, churn risk, average order value and overall lifetime value. Let’s take a closer look at the latter…

Using customer segmentation to drive higher lifetime value

The traditional metrics that matter when it comes to customer segmentation — those that you’d consider when determining whether or not your segmentation strategy is working — would include basic engagement metrics such as opens, clicks and conversions.

However, truly personalized segmentation needs to take this one step further. One of the most critical metrics to look at is lifetime value (LTV). High LTV customers are those customers that are predicted to be worth the most to you throughout the course of their time with you. Typically, the happier the customer, the higher the LTV. Without taking this information into consideration when creating your segments, you could be missing out on the full picture…

Let’s consider the above example. A marketing team might look at some initial segments based on demographic data around age and gender, and also factor in some value-based segmentation around average order value (AOV). With this view, considering who to target with your next campaign may look like a no brainer, with females aged 41-60 clearly spending more.

However, in reality, this demographic will only buy three times a year from your brand — maybe gifts for their teenage sons! — and without taking into account lifetime value you wouldn’t know that 18-25 males are more valuable to your business in the long run. Of course, this is something of a hypothetical example, but it does further emphasize the importance of taking into account as much data and as many factors as possible when segmenting your customers.

Gaining a holistic view of all your customer data to make these kind of well-informed segmentation decisions is, of course, easier said than done — let’s explore this in more detail below.

Data-driven customer segmentation: the role of artificial intelligence

According to Forrester research, 72% of businesses say managing data silos across multiple systems, technologies and regions to be moderately to extremely challenging. The lack of a ‘single customer view’ — a view that gives you that overall, holistic view of your customers, their preferences, their behaviors and more — can make combining your data and finding beneficial segments a difficult task.

Plus, even if you did manage to combine different data sets and uncover powerful segments, there’s then the question of which segments you should actually prioritize, which can be a hard decision to make even with a  clear view of all of your data.

Artificial intelligence (AI) gives marketing teams the ability to take your customer segmentation to the next level. As a technology, AI is incredibly well-suited to helping with segmentation due to its ability to make predictions and categorizations over large, complex datasets that a human alone would struggle to analyze.

It can spot patterns and identify trends that aren’t based on the more standard, basic segments (such as age and gender, for example). This empowers marketing teams to instantly identify groups of customers that are more likely to be highly engaged and valuable.

Let’s look at an example. Say you segment customers based on their propensity to purchase, so you know that you’re only speaking to those who are actively in the market to buy from your brand. This is handy to know, of course, but you need to do more; you need to also know what it is that they’re interested in buying, as well as their preferred channel.

You can do this by employing an AI recommender in each communication and touchpoint, enabling you to provide personalized messages at scale and build a model to determine their channel preferences. It’s all about identifying your customers’ previous purchases and behaviors, and using these to map future communications.

Say if you know that a customer has recently purchased a new pair of sneakers. A good example of going that one step further with your segmentation could be a follow-up communication introducing sneaker protection spray: “keep them looking white for longer!”

Introducing headless segmentation: a new approach to marketing

I hope that the steps covered in this article have got you thinking a little bit more about the potential that effective customer segmentation can carry for your business, and the key role that AI could play in delivering some game-changing enhancements to your existing segmentation strategy.

There may be a chance that you’re already at the top of your game when it comes to segmentation, and that you’re already doing much (or all!) of the above. But you might not yet be familiar with the idea of headless segmentation, a new approach to marketing that we’re pioneering here at Peak.

Headless segmentation is customer segmentation that doesn’t deal exclusively with a single system, platform or channel. But why is this so important? 

It means it can feed in customer data from multiple different sources, creating new customer segments that give you that all-important single customer view; their likes, dislikes, preferred colors, shopping habits, likelihood to churn and much, much more.

By gaining this new super-powered view, you can create more sophisticated segments — really, really useful buckets of your customers. And they don’t just have to be segments based off of purely historical data, either — headless segmentation, when coupled with the unique predictive abilities of AI, creates segments of your most valuable customers and predicts what they’ll do next. What’s the next product they’ll love, given what they just bought? When are they most likely to convert from an email send? This allows you to find and target more of the right type of customer — people that may have otherwise have slipped through the net of your enormous datasets!

Our on-demand webinar goes into this exciting new concept in more detail, outlining why we believe headless segmentation is the only way to win with your customers in the modern retail world. We also introduce you to Audiences, Peak’s AI-powered application that enables brands to deliver personalized experiences at scale. Here’s a quick overview of the application to give you a flavor of its capabilities ?

Remember, 52% of customers expect all of the offers they receive to be personalized — and that’s just today. Those expectations will continue to rise and it’s up to businesses to take action to future-proof their marketing by thinking differently about their approach to segmentation. AI-powered headless segmentation, we believe, will be the only way to deliver personalized experiences to every customer, at scale, all the time. 

…better get watching!