Data-Informed
How to segment your data effectively
So, you’re on the road to being a data-driven company. You’re tracking the metrics that matter for your feature or product. You’ve got visibility into the entire customer journey, and you’re baselining your performance. So what next?
Once you’ve got your baseline metrics – whether it’s signups, activation, conversion rates, etc. – it’s time to make them actionable. But how? A good place to start is segmenting your data. With smaller pockets of data to focus on, you can get a clear picture of what’s working well and what isn’t on a more granular level. You can then compare that with your expectations, goals, and industry benchmarks.
In this post, we’ll do a deep dive into the two phases of data segmentation: creating your dashboard and assessing user characteristics. We’ll walk you through each phase in detail. Let’s get started.
Why is segmenting your users important?
To drive business critical outcomes, you need to clear the noise of “average performance metrics'' and identify specific opportunities for removing friction or providing more value. You can do this through segmenting by usage, personas, jobs-to-be-done, verticals, channels, cohorts of signup or activation, and more.
Segmenting your users means you can see which group of users is experiencing a common friction point or value within your product. With this understanding, you can inform your roadmap and go-to-market strategy so that you focus on improvements to remove these common friction points for similar users. Equally, you can also help to acquire more of the users who experience value in your product.
Let’s look at an example. Imagine you want to look at the number of daily free trial signups on your website. You see that the average number is 35. Sure, this is helpful. But it’s not super valuable data that you can action any meaningful change with.
However, when you segment your data by medium, the detail in the picture changes. In the next example, you can quickly see that one label (in this case, “no-data" which represents organic traffic) represents the majority of the signups. The second best-performing source is PPC, while the others all show a negligible impact on this metric.
By grouping your users this way, you now know which channels are performing the best and the worst, as well as those that are sitting in the middle.
Understanding how different segments perform can help you calibrate your strategy, budget, and efforts around the initiatives you’ll put in place to drive more signups. It can inform you which experiments are worth pursuing and which ones you need to drop.
Now let’s look at the two key phases to segment your users.
Phase 1: Identify segments, baseline performance, and follow trends
It's important to understand that not all segments are created equal. Some segments are based on demographics, firmographics, or even where a user comes from. These are known as "acquisition segments,” and are essential for your marketing and go-to-market teams. These segments help folks make data-driven decisions that increase your customer base, decrease acquisition costs, and improve conversion rates across the funnel.
But there's another type of segmentation that's just as important: behavioral segmentation. This approach groups users based on their actions and engagement with your product or website. It's a game-changer for self-serve customer journeys or product-led growth teams.
By identifying usage patterns and understanding what makes certain groups of users successful, you can develop hypotheses that can be tested through experimentation. That could mean changing the digital experience, value or pricing, messaging, or even self-serve enablement.
While the first type of segmentation is relatively straightforward, many teams struggle to gain a clear picture of their behavioral segments. Follow the below steps which dive into both types of segmentation, with a focus on the second type. By the end, you'll have a better understanding of how to leverage behavioral segmentation to optimize your customer journey.
The attached worksheet walks you through each step and gives a repeatable framework. Feel free to go through that worksheet yourself, or distribute it to your team. We’ve provided a bit of a breakdown for you below, too.
Set up a working session with your team
Set aside a few hours for your team to have a working session. The process of this session should be as follows:
Step 1: Define customer journey “success” milestones (15-20 minutes)
By “success” we mean a specific event or set of events, that counts as completing the task your site, product, or feature is designed for. In some situations, this will be easy to figure out. In others, it may be more complicated. Some examples are: submitting a signup form, setting up the product for initial use, adding to cart, completing purchase, inviting a teammate, and so on.
Step 2: List the primary segments that apply to your business (15-20 minutes)
Below is a list of common user segments across industries and user types. Feel free to use these as a guide, or to choose other segments that are more relevant for the business results you or your team are interested in.
Binary:
Sometimes, if a user doesn't complete a certain step in the customer journey, it can help the team figure out what to do next to help them move forward.
For example, if someone starts filling out a signup form but doesn't finish it, the team can measure how many people usually finish the form, and then send an email to the people who didn't complete it with clear instructions and a video to help them out.
By doing this, the team can figure out why users might be struggling - maybe they need more motivation, or maybe the process is too hard. This can help the team investigate further, make improvements and measure their impact.
Frequency:
Some groups of users find certain features or products more valuable than others. By identifying these groups, you can learn more about what's working well and what's not.
You can also use this information to create new pricing or packaging offers for different groups or add new features that will be especially helpful for your most engaged users.
Step 3: Note hypotheses about your users (10-15 minutes)
Make a list of the top 3 characteristics of successful users of your product or site by analyzing user demographics, roles, or behavior. This process can help identify different assumptions within your team and generate hypotheses to test using data.
Step 4: Test your hypotheses (30-45 minutes)
Go through your product analytics platform data and see which hypotheses are true and which aren’t. For example, segment your users to find out if people from the West Coast buy more than people from the East Coast, or if users who check their balance once a day makes more trades than users who check their balance once a week.
Step 5: Discuss your learnings (15 minutes)
It can help to go around the room (or the Zoom) and see what comes up for people. What met your expectations? What didn’t?
Step 6: Create your dashboard for easy monitoring
Create a dashboard to show the success rate for different journey areas and segments over time. You can then break that down by the segments you’ve chosen in the previous step. You can check in weekly, monthly, or daily, depending on the needs of your business.
Step 7: Follow trends
Your final step is to see if the segments you’ve created perform the way you expected they would. Start reporting on them. Let your team know which groups need attention.
Phase 2: Discover and analyze new patterns in your users’ behavior
Your dashboards are an essential tool for tracking the impact of your go-to-market efforts and product or site experience on users during their customer journey. However, the agile nature of product delivery and growth initiatives means that things are always changing. As your team ships new capabilities and experiments, your dashboard will reveal new trends and friction points that require deeper analysis and potential changes.
It’s time for another team working session
If you notice poor performance or trends in key customer journey milestones, or with a new feature or experience, it's time to gather your team and investigate. Ask questions to understand what’s happening, where it’s happening, and why. For new features or experiences, analyze your most successful users and what they have in common, as well as your least successful users and their similarities.
The process of this session should be as follows:
Step 1: Start exploring (20 minutes)
Go through the list of successful events on your worksheet and circle some potential characteristics to investigate. Depending on your business, these may be demographic, firmographic, behavioral, or something else. Make sure to look at both qualitative and quantitative data–such as session replay or heatmaps. This will help you get the full picture of what happened, and why.
Step 2: “Scoop your neighbor” (20 minutes)
Now it's time for the fun part! Give everyone on the team 20 minutes to dive into the data and uncover some interesting findings. Feel free to get creative with the list of things to test - for example, do successful users engage with new features or use your in-app chat? And what about unsuccessful users - what common behaviors do they have? Don't forget to watch session replays or heatmaps to really understand what's causing friction or drop-offs.
Once you've made lists of shared characteristics and behaviors, it's time to deduce similar ideas and organize them into learnings and potential tests. The deliverable from this exercise should be a clear list of insights and action items that will help you improve the customer experience.
Step 3: Prioritize next steps (30 minutes)
Now it's time to vote for the top 1-2 things to focus on in order to improve the customer journey and boost business results. Prioritizing your product and GTM investments is both an art and a science.
Here are some simple questions to think about to start assessing the impact of your tests quickly and easily:
Which group of users is my experiment targeting? (casual, core, power, free, paid, etc.)
Is it going to degrade some users’ experience? How? How much? For how long?
What is the size of the test group?
Which usage/monetization metric(s) do we intend to move and by how much?
Would multiplying forecast improvement by the size of the group - yield the right/enough business impact?
How would this serve my short and long-term business goals?
Would this impact and tap into additional strategies and tactics needed by the company? For example - will this help renew 4 existing large customers, and at the same time open a whole new Total Addressable Market (TAM) that we can pursue to bring in new logos and grow?
How quick and easy is this to test? Do I have enough volume of traffic or a large enough install base to learn what I want to learn in a reasonable time? How much time will it take?
What does the opportunity cost? (alternative experiments we won’t be able to run)
–
It’s clear that grouping your users according to significant characteristics, behaviors, and other attributes gives you a clearer and more detailed picture of where and how you can improve the customer journey. Use it to inform your Marketing and GTM teams, messaging, positioning, channels, website, and signup flows - as well as in-app journeys or shopping experiences.
If you’re interested in putting these learnings into practice, then don’t stop at this blog post. Download our segmentation worksheets so you and your team can start building segments with ease. Don’t miss it!