Establishing a set of customer segments is the first step to strategic, impactful marketing. Beyond personalization, segments can be used to manage churn, increase margin, optimize engagement, and even mitigate fraud. If you are just starting out or want to get more advanced with your use of segments, read on to gain some valuable tips!
Man-Made vs. Machine Learning Segments
First, let’s review two types of segments.
“Man-Made” segments use historical data or existing information to group similar customers. Males who have purchased women’s jewelry in the month of February is one example.
Clutch offers another option called Clutch Powered Segments. These segments use machine learning algorithms that predict what your customers are likely to do in the future, and are more accurate than humans in identifying the right customers for each group. Using the above example, Males who are likely to purchase jewelry in the month of February would be the equivalent Clutch Powered Segment.
The difference is that Clutch Powered Segments continually learn from customer data (like transactions or historical behaviors) and adapt to your brand’s specific customer behavior, making them more accurate and more relevant over time.
How many segments should you have and which traits should determine a segment?
This depends a lot on resources, but starting with 3-5 segments and working up to our top 10 (described below) is feasible for most mid-sized brands. Just remember that the more segments you have, the more content you need. So scaling up the number of segments can mean adding support from an agency or content management platform if you plan to leverage segmentation for multiple touch points.
A good approach is to focus on one goal and work backwards from there to determine the segments needed. To “Identify my top customers” you might start with one segment: Customers with X total lifetime spend and X average visits per month. Later on, you could create additional “Top Customer” segments to identify top customers who are big spenders but less frequent shoppers, and so on.
What are Clutch’s Top 10 Strategic Segments and what data is needed to create them?
1. Top Customers
-
- Definition: A customer is a Top Customer if they are within their predicted future lifetime value is among the top 10% of all customers
- Data Needed: Transaction data – dates, amounts
2. Active Customers
-
- Definition: A customer is Active when they make a purchase. The customer remains Active until their time between purchases is longer than 85% of their previous time between purchases.
- Data Needed: Transaction data – dates, amounts
3. Customers At Risk of Churning
-
- Definition: A customer is At Risk when their time between purchases is longer than 85% of their previous time between purchases.
- Data Needed: Transaction data – – dates, amounts
4. Churned Customers
-
- Definition: A customer is considered Churned when the time between purchases exceeds their longest time between purchases.
- Data Needed: Transaction data – dates, amounts
5. Future Lifetime Value
-
- Definition: This segment can be made up of all customers or a subset, and represents the predicted value of a customer based on recency, frequency, and monetary spend.
- Data Needed: Transaction data – dates, amounts
6. Possible Fraudulent Accounts
-
- Definition: Customers accounts are flagged as fraudulent when they demonstrate transactional and behavioral activity that is suspicious or outside the norm as defined by the brand.
- Data Needed: Transaction data (dates, amounts) and brand feedback
7. Optimal Purchase Channel (online or in store)
-
- Definition: These segments are made up of customers who prefer to make purchases online or in store
- Data Needed: Transaction data including whether the purchase was in-store or online
8. Optimal Communication Channel (email or text message)
-
- Definition: These segments are made up of customers who prefer to engage on email or via text
- Data Needed: Open and click through rates for communication channels by customer
9. Optimal SMS Send Time
-
- Definition: This segment indicates the best time to send a mobile message for optimal engagement
- Data Needed: Open and click through times for SMS by customer
10. Optimal Email Send Time
-
- Definition: This segment indicates the best time to send an email for optimal engagement
- Data Needed: Open and click through times for email by customer
** BONUS SEGMENT** Lookalike Audience
-
- Definition: This segment identifies customers who display similar behaviors to self-identified or known customers
- Data Needed: A known data point (ex: female) and transaction dates and amounts.
How can segments be used to achieve more advanced marketing objectives?
Top Customers | Your most valuable customers
As a marketer you know it’s important to stay connected with your Top Customers since they provide the most long-term value. Continuous touch points with these customers can make them feel valued, increase their engagement and ultimately, increase their spend. The Top Customers Segment includes your top 10% of customers who have the highest predicted value over their lifetime. This segment can be used individually, as a starting point to build your own segment, or as a part of customer journey mapping and event triggering.
Customer Lifecycle | Customer Lifecycle is split into three segments that follow the customer journey path – Active, At-Risk, and Churned Customers
As a marketer you understand it is a lot more expensive (studies have shown 5x more) to acquire a new customer than to retain an existing one. The million dollar question is how do you keep an Active Customer active and prevent them from ever reaching a churned state? The answer is by focusing your targeting efforts towards Active and At-Risk Customers.
- Active Customers are your customers who are likely to make another purchase. If you have limited resources and want to ensure your budget is going as far as it can, this is a good go-to group to target. You’re speaking to customers who are actively engaged with your brand and are more likely to get you a return on your investment.
- At-Risk Customers are your customers who are still active, possibly high value, but likely to churn soon. This segment is important to act on immediately because it provides a limited window of time where your customer needs an incentive to shop with your brand again. The At-Risk Customers segment can be tied to a triggered campaign where as soon as a customer falls into an At-Risk state, they are targeted with rich and relevant offers to stay engaged and build loyalty back up. Using the At-Risk Segment and Triggered Campaign strategy eliminates the need for an at-risk customer analysis, which is usually stale and too late to act on by the time it’s delivered.
- Churned Customers are your customers who are not likely not to make another purchase. Unfortunately, no matter how hard you as a marketer tries, there will always be a population of churning customers. However, the more we focus our efforts on Active and At-Risk Customer segments, the less we’ll see customers fall into the Churned Customer Segment having had no marketing touch points. If you want to target this group separately, you can target them with a winback campaign. The most valuable effort is to analyze common characteristics to pinpoint additional opportunities for bringing these customers back in. Do a lot of recently identified churned customers live near a competitor that recently popped up? Have a lot of these customers left bad reviews prior to churning? There’s a lot a marketer can gain from learning about this segment of customers.
Optimal Purchase Channel | In Store & Online Shoppers – Customers who make most of their purchases in-store versus online
As a marketer you engage and cater to your in-store and online customers in completely different ways since they shop in completely different ways! The In-Store Shoppers and the Online Shoppers Segments identifies shoppers preferred shopping experience. These segments can be used to tailor your marketing engagement for their preferred shopping method or even educate and increase exposure on their less preferred shopping method.
Optimal Communication Channel (email or text message) and Optimal Send Time | The communication channel and time of day with the highest engagement for an individual customer
As a marketer you always have the goal to increase customer engagement with your communications. One easy way to do that is by increasing the relevance of how and when you speak to them. Optimal Communication Channel identifies the communication channel the customer is most likely to engage with while Optimal Send Time identifies the time a customer is most likely to engage. When you pair these two together you’re providing your customer with a totally personalized experience and will naturally increase your communication relevancy.
Possible Fraudulent Accounts | Customer accounts that have suspicious loyalty activity that could be deemed fraudulent
Loss Prevention (LP) teams know that loyalty programs are easy targets for customers and employees looking to game the system. Today, LP teams are manually identifying suspicious customers by arbitrary business rules but may unintentionally flag their most loyal customers instead. The Possible Fraudulent Accounts segment identifies the most suspicious customer accounts who may be committing fraudulent activity. LP teams can improve the segment by providing feedback on how they define fraud so we can tailor the segment to their brand. With this segment and new efficiency, they can cut out the legwork, prioritize customer investigations and shut down fraudulent accounts.
*BONUS Segment*
Lookalike Audience | Customers who display similar behaviors to known customers who have self-identified. Ex: Customers with specific dietary preferences – vegetarian, vegan, gluten free, paleo, etc.
Today people have adopted many specific dietary preferences, so collecting this data is a great way to personalize content in the food service and hospitality industries. But how do you cater to most of your customers who haven’t provided that information? Using a Lookalike Audience, we can find additional customers who display similar behaviors to those individuals who have already provided their dietary preferences.
We can use engagement as a metric to confirm our findings. For example we can send vegan-friendly recipes to our Vegan Lookalike Audience and measure engagement to see if our audience is accurate. Furthermore, we can use this opportunity to collect additional data points, such as whether they clicked on the Asian inspired or Mexican recipes.
The utility of segmentation goes far beyond traditional marketing. They are the foundation of many business intelligence strategies, revenue forecasting and growth efforts. Using segments effectively can help you understand your customers and how to interact with them, driving much of the impact you will have as a marketer.
