Creating customer segments is all about dividing your audience into smaller groups based on things they have in common. Naturally, you can do this by age, location, gender and other demographic classifications. This is only the beginning, however.
In this article, we’re going to look at some of the ways you can combine different data types to create highly targeted and effective segments that deliver results.
Customer segments need data!
First things first, the raw material needed for customer segments is data. Depending on the purpose of the segment, you might require a combination of several different data-points. Below are the four most common categories of customer data used. But do recognise that any piece of information you have about a customer should be usable.
Here we are looking at things like age, nationality, gender and potentially something like income level. Many purchase decisions are influenced by these factors, but be careful of how heavily you lean on them in communication. Be careful of being “too personal” with your content. It’s ok to use these data-points as a condition of the segment, but that doesn’t mean you need to call attention to them in your actual copy and creatives!
What actions has the customer taken, or not taken? Have they made a purchase recently, and what was it? Or have they visited your site or opened your app a certain number of times in the last week? Any action, or a lack of action, can be a condition for inclusion in a segment.
Also known as zero party data, this refers to your customers’ preferences, opinions and intentions. This type of information is incredibly valuable and is often collected through interactive surveys or via on-site and in-app preferences centres. We recommend you check out our blog to learn more about this!
Clearly this has to do with a user’s physical location, but it extends beyond the country or even the town. iBeacons and geofencing technology allows brands to contact app users based on their exact location, e.g. in a particular store or a stadium. You might want to create a segment of app users who are currently within walking distance of a retail outlet to promote a flash sale, for example.
Important customer segment terms
We’ve boiled the topic of segmentation down into three key concepts, which are important to understand.
Note: Don’t worry if this is confusing, as it will all make more sense when we look at how they all relate to one another in real-life campaign examples!
These are the qualifying reasons why a customer is added to a particular segment. The range of conditions available for selection within the Xtremepush platform is huge, and includes everything from the type of device the customer uses, to the source of attribution (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 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 makes a successful customer segment?
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/ Goal-Driven
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.
- Multiple conditions
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 most common use cases for a customer segment?
There’s a near-infinite 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.
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.
As a little extra, we are also making their preference for web push a condition, as this is the channel we’ve chosen.
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 VIP customers (determined by the corresponding attribute) 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 VIP 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.
Too much of a good thing?
There is one potential risk that we need to be conscious of with all of this great segmentation; over-messaging. It is possible that a customer may inadvertently fall into several segments and find themselves enrolled in multiple campaigns or workflows in a short period of time. This can happen if a customer hits a number of triggers and you have numerous attributes associated with their profile.
However, on the Xtremepush platform, we have a concept of messaging limits, which basically allows you to set a strict limit on how many messages (across all channels) a customer can be sent per day, per week and per month.
In the example below, we have created a category of message, “Promotional”. When creating a campaign, you can insist that a category is picked. For this type of message, we have put clear limits in place, to safeguard against over-messaging and potentially damaging the relationship with a customer.
We encourage you to experiment with different combinations of conditions to identify the most valuable segments. And remember, it’s best to start at the desired end goal and work backwards.
If you have any further questions about customer segmentation, or what to discuss potential use cases then feel free to get in touch. Xtremepush is an award-winning multichannel engagement, personalisation and data platform that makes segmentation easy.