What is customer segmentation?
Customer segmentation is a way of grouping your entire contact list into smaller audiences based on shared attributes, behaviours and demographics.
The goal of customer segmentation is to improve the relevance and ultimately the performance of your campaigns, targeting specific audiences with content that reflects their interests and behaviour. The desired impact is better open rates, click-throughs and conversions at every stage of the customer journey.
There is also a significant cost-saving implication to good segmentation. By prioritising or identifying only the exact group of customers that you want to engage with, you are minimising the scale of your campaigns. This means fewer emails, SMS, push notifications, and so on, sent to customers who would rather not receive them.
And of course, hand in hand with this, the other benefit is an overall better customer experience, where the consumer feels valued and understood.
In this guide, we’ll be looking at the concept of customer segmentation from a variety of angles, focusing on the tried and tested approaches to creating segments that deliver results.
We’ll also look at some commonly used segments and where they fit into an overall customer acquisition, engagement and retention strategy.
More than anything, we want to emphasize the changes in segmentation methods, made possible by first-party data.
What you’ll see is that the idea of segments as fixed, unchanging lists has been replaced by dynamic, real-time audience creation at the point of campaign delivery.
The importance of data in customer segmentation
True segmentation relies on accurate first-party data, whether that’s information on the customer’s behaviour, spend or interests.
Looking at the image below you can see that segmentation typically takes place towards the end of the data management cycle, as you prepare your audience for activation.
One of the more recent developments in digital engagement has been the concept of a Single Customer View (SCV).
We recommend reading the article linked above for more detailed information on the SCV, but fundamentally it means consolidating all of the data you have collected on each individual customer.
What this means is that you can create much more accurate, in-depth segments which draw on all types of data.
Personalisation, or one to one marketing, comes in two forms; a) you can ensure the actual content of your campaigns is individually relevant (their names, account details, products they’ve bought, etc) and b) you only send them messages that are relevant to their interests or where they are in the life cycle.
And of course, you can always do both!
Segmentation is perhaps the easiest and most effective way of putting your data into action and personalising your marketing efforts.
Static vs Dynamic customer segmentation
What is static segmentation?
For a long time, the list was the lynchpin of segmentation. You created your list and then uploaded it in CSV format to whatever engagement solution (usually email) you were using for execution.
That particular list was fixed, or static, at the moment it was made.
For recurring campaigns, typically daily or weekly, a new list would need to be generated and uploaded each time.
If you’re using a CRM system then you’re going to be very familiar with static lists, and their limitations.
List-based segmentation is also associated with “batch and blast” campaigns that target a large percentage of your customers with a generic message. This is something that most organisations are trying to get away from.
And dynamic segmentation is the way forward.
What is dynamic segmentation?
A dynamic segment is fluid, populating in real-time whenever the campaign is sent.
It’s the basis of automated, real-time engagement, allowing campaigns to be triggered as and when customers enter or exit a segment.
For example, you might want to create a segment of first time buyers, with a view to sending them a personalised welcome message when the purchase is complete. But should a customer fail to make a second purchase within the next 3 months, let’s say, they would be moved into a new segment of lapsed or unengaged customers.
The specifics are very much dependent on your own goals, but the principle remains the same; customers naturally flow in and out of lists depending on how their relationship with you evolves over time.
Dynamic segmentation makes this easy to achieve, updating each audience grouping with real-time accuracy.
What’s the advantage of dynamic segmentation?
Of course, there is still very much a place for list-based segmentation. Marketers and CRM professionals use it alongside the dynamic method for a variety of reasons, but it is most common in B2B marketing, where leads are bunched together based on their source (registered for a webinar, attended an event and so on).
For B2C marketers, however, dynamic segmentation is generally viewed as the ideal solution, reflecting the ever-changing relationship that brands have with consumers.
And there are practical advantages, as well as strategic. Aside from the unnecessary time wasted creating and uploading static lists, you also run the risk of a contact no longer actually meeting the criteria (or conditions, more on that in a moment) for inclusion. For example, they may have opted out of marketing emails in the time between the upload and campaign execution.
It’s worth noting that it’s still possible to export a dynamic segment as a list if you desire.
Building blocks of dynamic customer segmentation
At the core of real-time, dynamic segmentation are three elements; conditions, events and attributes.
These are the qualifying reasons why a customer is included in a given segment. The range of conditions available for selection within the Xtremepush platform, for example, is huge, and includes everything from the type of device the customer uses, to the source of attribute (for app installs) to their general on-site and in-app behaviour.
More often than not you will be segmenting based on an event and/or an attribute; so that’s something the customer has done/not done and something to do with their personality. Below is an overview of the types of conditions available, and of course there are many deeper options within each.
One of the most important elements of segmentation is the idea of “attributes”. An attribute is a sort of tag that is assigned to a customer’s profile. It can be based on any number of factors, as decided by you, and a profile can have as many different attributes assigned to it as you want. These attributes in turn are then used to segment your users for micro-targeting.
Here are three examples of attributes that a customer might have associated with their profile, visible via the Single Customer View.
It’s also worth mentioning “events” too, as they are often the starting point for a campaign, or a condition for including/excluding customers from a segment. An event can be any clear action taken by a customer on your website or app. You may very well be using something like Google Tag Manager (GTM) at the moment to tag particular events in order to measure performance and calculate goal completions.
As one of a select number of certified GTM partner vendors, the Xtremepush platform allows you to directly import any existing tags for immediate use in campaigns and segmentation.
What is micro-customer segmentation?
Without putting too fine a point on it, a micro-segment is a highly targeted group. It combines multiple conditions, building very small groups of your users.
How small? We have many clients who create micro-segments with tens of customers in them, allowing for granular, laser-like engagement.
As a rule of thumb, the smaller the segment the more likely it is that the campaign will perform well.
Of course, that prediction is based on two things: a) through your combination of conditions, you’ve identified a cohort of customers who are need of or are likely to buy something quite specific and b) your creative assets (image, copy, offer etc) all speak to this.
How to create effective customer segments
The two common ingredients we see across all of our clients’ segments is that they take multiple data points to create highly targeted groups and they also have a very well-defined purpose in mind for each segment.
ROI and goals
This has got to be your starting point when creating a customer segment. Don’t waste time creating segments that are not going to help you drive your core business goals. The goal should be quantifiable in some way, whether that’s directly attributable revenue from a campaign, driving traffic to your website or re-engaging a lapsed customer. From here you can start to apply rules to identify the ideal group, before creating the actual campaign itself.
Having a clear goal in mind also a great way of identifying where there are gaps in your data. If you don’t have the right data in order to create the segment you need, then that’s something you need to address.
The most successful marketers create segments based on a number of conditions, using classic Boolean and/or rules. For example, you create a segment of users who haven’t opened the app in the past 2 months and are interested in soccer and are using an Android device.
This level of depth allows you to create hyper-personalised campaigns that resonate with the recipient and drive key business goals. The ability to layer multiple conditions in this way should be a non-negotiable when researching a vendor.
What are the typical customer segments?
Every brand will have its own unique challenges, and its own way of categorising customers. How you group your customers may be very different to another brand’s approach.
However, no matter what title you give each segment, here the 5 pretty much universally accepted archetypes.
Ok, no medals will be given for working how who belongs in this group! But there is slightly more nuance than just when they made their first purchase.
You may also want to factor in information like which type of product they bought, if it was during a sale, and where they came from (email campaigns, digital advertising etc).
By doing so you can create a more compelling onboarding or welcome journey, and ideally nudge them towards a second purchase.
These are customers who have recently made a handful of purchases and are showing signs of being long term supporters of your brand.
You may want to sign the, up to a membership scheme or loyalty programme that offers rewards and keeps them buying from you.
There will always be a certain number of customers who, despite early signs of being with you for the long haul, start to drop off the radar.
They visit your site less frequently than they used to and crucially, they are not spending as much as they once did.
We all know it’s more cost-effective to retain a customer than it is to win a new one, so it’s crucial to engage “at risk” customers with a personalised incentive to make sure they don’t disappear forever.
The creme de la creme, the golden cows, the VIPs. These are your high-priority customers who make regular purchases and spend plenty of money with you.
This is the ideal segment to offer early access to new products and services, with the goal of converting them into active advocates on social media.
This is absolutely one segment you want to stay in regular contact with.
And finally, we come to the churners. It’s not you, it’s them!
Despite your best efforts, some customers log out of their account one day and don’t come back.
Creating churned (or inactive) segment allows you to deliver a highly-targeted win back campaign which more than likely includes an offer or discount that you wouldn’t want going out to all of your customers.It’s a “last roll of the dice” scenario,
This is a principle of good data management in general, but after a significant amount of time has passed without any signs they might return, you’re better off deleting churned customers.
Any approach to segmentation is essentially just the exact way you go about creating some version of the 5 groups above.
In this guide, we’re going to focus on one method, in particular, RFM.
The R.F.M. approach to customer segmentation
This stands for recency, frequency and monetary.
Essentially, this is an approach to segmentation centred around a customer’s behaviour to date within these 3 broad parameters. In other words, it’s segmentation based on when a customer last did something, how many times they have done something, and how much they have spent or are worth to you.
When was the last time this customer made a purchase, visited your website, opened your app and so on. The answer is a very obvious indication of how top-of-mind you are to them.
A lack of recent activity would suggest that a re-engagement campaign should be sent to reduce the likelihood of them churning.
With a given time period, how often has the customer interacted with you? Analysis of frequency is the clearest indication of their loyalty level, ranging from devoted fans to one-time only drop-ins.
Defining segments based on the frequency of engagement means you can enroll customers in journeys designed to either nudge them to the next loyalty tier or ensure they get the VIP treatment they deserve.
Understanding how often a customer plays or buys means you don’t unnecessarily send them incentives like bonuses and discounts.
How much money has the customer actually spent? This could be a lifetime value, or again you may just assess it over a set period of time. If you could speak to your most profitable customers, what sort of message would you want to send them?
You might want to give them early access to new content or promote products ahead of their official release.
Or you might just want to say thanks and let them know how much you appreciate their business.
What are the most common uses of customer segments?
There’s a vast number of potential segments you can create, even with a relatively small set of data-points to use as conditions. But essentially, the most common and practical goals are to a) drive revenue or b) drive engagement.
In the examples below, you’ll see a mix of segments built around conditions, events, and a combination of both.
As we said earlier, it’s time to move away from the old idea of static lists. The modern approach to segmentation is far more fluid, centred on micro-behaviours and real-time actions.
Early lifecycle and onboarding
The early touches you have with each new customer very often dictate the course of the relationship.
It’s quite common to enroll a new customer in a dedicated onboarding or welcome journey. In this instance, the initial message in the sequence is typically triggered by a first purchase event.
This is another example of where dynamic segmentation comes into its own, allowing you to automatically move a customer from an “early stage” segment after a certain period of time has elapsed.
One of the most frequently used conditions for a segment relates to a customer’s loyalty level or total lifetime value.
The necessary customer data usually comes from a loyalty platform (e.g. Crowdtwist, which we have seamless integration with), or from a backend eCommerce or CRM solution. As in the example below, even using a single condition can help you to identify your high-value customers.
From here it’s really just a matter of strategy, identifying ways to further strengthen their bond or nudge them towards advocacy.
Abandoned cart recovery
There are several ways to set up an abandoned cart recovery campaign, but using segmentation is one of the most effective methods.
The first thing to note is that we have an event-trigger in place, “product_added_to_basket”. That’s what kicks off the campaign. And we are waiting 24 hours before the recovery message is sent.
You might ask at this point, what if the customer came back themselves and completed their purchase, would they still be sent the campaign? That’s where segmentation comes in. We make the event purchase_made a condition and check that it didn’t hit in the meantime.
For an extra layer of depth, we are also making their preference for web push a condition, as this is the channel we’ve chosen to send the campaign through.
And here’s how the campaign appears to the customers. You’ll notice that we’ve personalised the creative with both the name of the customer and an image that relates to the product added to the bag. This is an example of dynamic content in action.
Re-engage dormant users
In the next example, focused on the sports betting industry, we are concerned with high-value customers who have been worryingly inactive over the past month. The implication here is that they are on the verge of churning completely, so we want to intervene before that happens.
We don’t have a specific event triggering this campaign, but we are using the event bet_placed as a condition and checking how often they have done so in the past month. This kind of campaign could be automated to run on the first day of every month, or you might prefer it manually at regular intervals throughout the year.
In the publishing and media industry, we would often see brands using website_visited as the conditional event. Again, the freedom is there in our segmentation engine to operate however is best for you.
So, we have now identified a segment of customers that we very much want to re-engage. From here we can create a relevant campaign that speaks to them in a particular way and perhaps even includes an incentive to come back.
And the list goes on!
There’s no limit to the kinds of segments and use cases that are possible, provided you have the necessary data to use as conditions for it.
Get in Touch
Xtremepush is the world’s leading customer engagement, data, and personalisation platform. We enable enterprise brands with ambitious goals to collect, unify and activate their customer data across the most important digital channels.
Our powerful segmentation engine allows you to create granular audiences, leading to increased revenues and better customer experiences.
Contact us today and learn more about how we can help you connect with your customers.