While ferocious Nazis make for a better film, the US Air Force also faced a lesser-known enemy in the 1940s: air accidents. In 1944, a typical year, the rate of fatal accidents was 6 per 100,000 flying hours - 150x higher than modern aviation.
A study investigating the causes made a surprising discovery. Cockpits were designed to fit the average man. Yet after measuring over 4,000 pilots, not a single one fit the average in every dimension.
There was no average pilot. Cockpits designed to fit the average pilot actually fit nobody.
This same trap, the myth of the average user, is easy to fall into when building a product. Building for just one type of user risks alienating huge parts of your userbase.
User segmentation lets you understand who the different groups of people using your product are. I'll stop short of calling it a superpower, which is joining the likes of ninja and rockstar as a startup cliché.
Even so, it brings a deep user understanding. As the case studies below show, you can leverage this to drive better results across your business.
There are many types of segmentation and goals it can help you achieve, so it's hard to know where to begin.
In this post I'll cover how to segment users in four ways, and how early-stage SaaS businesses can take advantage of these to:
User segmentation is dividing your users into groups based on their characteristics or behaviours. There are different segmentation types:
Which type of segmentation to use depends on the goal you are trying to achieve. Below I'll cover the type needed to achieve each goal.
Before we start: don't go narrow with your segments. The point of segmentation is to understand your main groups of users, not to cover every last edge case.
Why? If you end up with 14 segments, you're not going to be able to make changes for each of them. Aim for 2 to 4 segments as a general rule of thumb.
Churn can be a killer for a subscription business. Even low levels of churn, such as 5%, can make a big dent in your long term growth prospects.
Looking at how users behave lets you see which actions (or lack of actions) are linked to users churning shortly after. This knowledge of which users are a high churn risk is powerful. It gives you a chance to re-engage them and improve retention.
The team at VWO (an A/B testing tool) segment users into engagement buckets. These are based on how many times users take certain key actions in the app over a set period.
Users in the low and critical segments are at high risk of churn, and are entered into a personalised email flow to re-engage them. This proactive approach hugely increased VWO's retention (full case study).
How to replicate this
1. Use the following data to create a 'low engagement' segment. Play with the parameters till you have a segment that correlates with churn, meaning: a) most users who did churn would have been in the low engagement bucket and b) few users who did not churn would have.
2. Create an email (or email series) for re-engaging the 'low engagement' segment of users. It can include new features, blog posts, case studies, features they may have missed and more. It's also a great opportunity to ask if they have any questions or need any help. This article has email template ideas.
Just please, don't beg. It's lazy and it won't work (but if it does, let me know).
3. If you have a large userbase then automate these emails. If you have fewer users you could go manual and add personalisation.
It won't scale but users will appreciate the personal touch. Plus it will create a more natural path to discuss why your product may not be meeting their needs.
Behaviour can be a warning light that a user is about to leave, but there's a 'glass half full' side to it too. Behavioural data offers clues that a user is having a great experience with your product. To maximise conversions, your key goal in onboarding should be getting users to the Aha! moment as soon as possible:
“The initial Aha! moment is the moment a user realizes value in the product, or is convinced that it will be a solution for a specific pain they have." - Joe at Userpilot
Different types of users may have different actions which lead to their Aha! moment. The goal is to find out 1) which actions are linked to a higher conversion rate, and 2) how that differs based on user groups. You can then encourage more of those actions and increase conversions.
MYOB is an Australian accounting SaaS. They initially offered the same trial experience to all users. Then behavioural segmentation showed them different types of businesses had different Aha! moments.
Shops, for example, were more likely to convert if they'd used the point of sale retail features. Factory users cared more about features that made their payroll process more efficient.
Off the back of these insights, MYOB moved to a tailored trial. This helped more users find their Aha! moments and increased trial conversion rates by a whopping 54%!
How to replicate this
Are you already tracking user behaviour? If you aren't but it interests you, this guide has a good overview of your options.
If you aren't interested in behaviour tracking skip to step 2.
1. List the behaviours displayed by most users who convert from a trial, and are rarely shown by users who churn. What to look at depends on your product, but here are a few ideas:
2. Talk to users who have converted. If you have behavioural data this will add context and provide the 'why' behind the behaviour. If you don't have behavioural data, this will be your main source of insights. So keep speaking to more users until you're not hearing anything new. Ask about:
3. Make changes to your onboarding flow to get users to Aha! moments as fast as possible:
Some SaaS businesses prioritise new features based only on the number of users who vote for them. This is the myth of the average user in action.
Behind the votes are different groups of users. If you don't account for them you could build a product that doesn't meet the needs of any group. Or you might build features that only benefit users on free plans while ignoring paying customers.
That's why leading product management tools such as Productboard and Canny let you segment by user when looking at feature requests (and no, those aren't affiliate links. I'm pretty sure this blog only has 5 readers anyway).
Say the Evernote team only has time to build two new features. If they pick the two with the most votes, they won't have satisfied either corporate or personal users:
If you understand your user segments' needs, you can account for whose needs you are meeting when prioritising features. You may also want to put a value on each segment (number of users x LTV). This will let you prioritise the needs of your most valuable user group.
How to replicate this
1. Set up short, casual chats with users to find out:
2. Continue speaking to users until you can identify groups who share similar jobs and needs.
3. Now you have the segments, create and send a survey to users you haven't spoken to. Ask about their jobs/needs and segment accordingly.
You can also look at product usage to see different user groups' needs. For example, a travel booking app may find from speaking to users that anyone searching for hostel stays of 2+ weeks tends to be a backpacker. This helps to calculate the size of the backpacker segment.
Marketing that's relevant to its audience performs better. Which makes sense. You're more likely to click an ad for that getaway you've been eyeing than another ad from the latest venture-backed grocery delivery app (sorry Getir but I must've seen it 30 times over).
The first step is understanding who your target users are. IT services provider Marco was struggling to bring in blog traffic. So they overhauled their blog with a new focus on their target users.
Using demographic, psychographic and geographic segmentation they built several user personas. Each one represents a different segment of their users. Importantly, they identified the needs of each segment as well.
They went from general tech posts to posts that directly spoke to their target user's needs. Huge success followed, with traffic to their new blog jumping from 88 visits per month to 2,249 visits per month (a 2,500% increase) in the space of one year.
How to replicate this
This strategy is not limited to blog posts. You can tailor ads, cold emails, social media posts and any type of marketing to speak to a certain audience.
For more info on creating personas Neil Patel has a good guide.
Bonus tip: a 'quick win' from segmentation is to segment users based on which marketing channel they come from. Then calculate the LTV of users from each channel, so you can double down on the strongest one.
Your existing users are a goldmine of information about who wants your product and why.
Insights from segmentation let you take actions that can transform the fundamentals of your business. As the case studies above show, there are huge wins to be had. Plus they compound over time: imagine the impact a 20% increase in conversion rate would have on the long-term growth of your SaaS.
Don't wait until you have a bigger team to start segmenting users. Better user understanding will enable you to start making smarter decisions right now about how to build and market your product. SaaS consultants found most of their successful clients use multiple types of segmentation.
So start now. For early-stage SaaS businesses, segmenting users isn't something to leave until you grow bigger. It's a tool to help you get there.