The Secret to High Retention: Your Product Must Be Loveable!

Using user segmentation to identify areas of delight

Best Practices
July 26, 2023
Headshot of Jiaona Zhang (JZ)
Jiaona Zhang (JZ)
SVP of Product at Webflow
Reforge Blog Series blue, purple and teal header image in Amplitude branding with headshot featuring Jiaona Zhang

This blog is part 2 of a 3-part Reforge blog takeover series, where Reforge experts help product and growth leaders improve their retention. Like JZ’s perspective and looking for more? Be sure to check out her Reforge courses on Product Management Foundations and Managing Your PM Career.

One of my core beliefs as a Product Leader is to build loveable products. Too many PMs launch things that don't always meet the “lovable bar.”

Contrary to the somewhat broken industry standard of minimum viable product (MVP), a minimum loveable product (MLP) goes beyond functionality or usefulness. Instead, it demonstrates an acute understanding of what users find valuable in order to ultimately delight your users.

I use this analogy with my students at Stanford: If you’re trying to test whether or not people like pizza and you serve them burnt pizza, you’re not going to find out if they like pizza, only that they don’t like burnt pizza.

At Webflow, a viable product for our audience has the basic functionality in place for users to create a custom website. The product becomes loveable when we add in the elements that make the design experience seamless and delightful, like enabling keyboard shortcuts for tedious actions, taking complex interactions and making them possible without code, and radically simplifying the editor experience.

At face value, building products people love sounds like a great idea. But deeply understanding your users enough to build a lovable product and continuing to evolve and deepen that understanding to retain users can be challenging.

There are usually many types of users at any given time, and as a product manager, you’re building for both current and future users across many demographics and use cases. One of the best tools in addressing these problems—building a lovable product and keeping it lovable for different types of users—is user segmentation.

If you understand who your users are, then you can build a loveable product that is targeted at each user effectively.

Different features are “lovable” to different parts of your user base

Many products have a broad, diverse target audience or user base. But even if everyone in your target audience experiences the same problem, they often have different needs in an activation experience, different vantage points on what a “good” experience feels like, and different barriers to building a habit with your product.

Let’s consider Airbnb as an example. Airbnb is a loveable product for guests to find incredible places to stay around the world, and for hosts to earn passive income on unused properties. Airbnb has a wide range of audience segments who host on their platform, from all different parts of the world, and who use Airbnb for different reasons.

The product experience for property managers who list hundreds of properties at once looks very different from an individual looking for passive income from their vacation home while they’re away. To improve retention in the property manager segment, the Airbnb team needed to build additional features for connecting to existing property management systems, calendaring tools to bulk edit listings, and mass messaging tools.

In contrast, to retain more individual hosts, Airbnb needed to remove complexity from the product experience to make it extremely easy for individual hosts to list their property on specific dates, reply to messages, and set the right price.

This tension is also present in B2B products, like Webflow. Webflow is beloved by users because it’s powerful and fast. You can use Webflow for so many different use cases, from independent small businesses—like dentist offices and coffee shops that want an easy-to-use tool—to enterprises like Nike and Amazon that use Webflow to reduce their development costs.

There may be users in both segments that won’t retain and say it’s because Webflow has been hard for them to learn. Yet the solution that the team should implement to improve retention is different for each segment. For an independent use case, the team might explore creating a more streamlined experience with less customization upfront and more templates to streamline creation. In contrast, for an enterprise use case, they may explore routing those users to get additional support from a third-party service provider.

In both situations, it’s critical for the teams to understand which segments are using their product, how they’re using the product differently, and build loveable moments for each accordingly.

Building a lovable product that retains users starts with better user segmentation

So how do product managers navigate the tension inherent in building for many different types of users and broad audiences? They dig into user segmentation.

User segmentation is the process of dividing up users and grouping them. But more specifically, user segmentation is looking at how a specific population behaves differently from another or contributes to a metric over a period of time. User segmentation is the key to driving true value for your users and your business. User segmentation is an incredible tool for finding insights that can lead to new feature ideas and better retention.

Although segmentation can be a powerful tool for product managers working on retention, they face three common pitfalls:

  1. Not segmenting
  2. Using poorly defined segments
  3. Over-indexing on churned users

Segmentation pitfall 1: Not segmenting

One of the most common and most painful pitfalls when product managers work on retention is not segmenting their user base or fully understanding the value of segmentation.

Although segmentation is a common tool for evaluating top-of-funnel patterns, product managers often overlook important segments in their user base when evaluating retention and exploring opportunities to improve retention. This often looks like plotting a retention curve or analyzing engagement behaviors across the user base. This treats your users as one big blob and often results in drawing insights from the average.

When you only consider the aggregate, or the average, you miss critical insights on ways to improve, or worse: You assume something is going well when it isn’t. Further, good segmentation can improve your user research for future features or product development to ensure you’re getting feedback from the right people.

When you do user segmentation well, you mitigate the risk of missing insights based on the average. But applying segmentation is just the first step; it’s also important to ensure that the segments you’re using are well-defined.

Segmentation pitfall 2: Using poorly defined segments

User segments should be defined in a way that’s meaningful for the problem you are trying to solve.

One of the most common ways product managers end up with poorly defined segments is by relying on self-reported information. Early on, many teams will use surveys or ask for information in a user profile to inform user segmentation. Although this can be a scrappy way to understand differences in who is using your product, self-reported segmentation information always has limitations. Although you may think your survey questions are clear, this wasn’t the case at Webflow. Here are the two challenges we ran into:

  1. Users often reply in ways inconsistent with how you think they should be answering the questions
  2. Individual users change over time, but surveys only capture a moment in time

At Webflow, we frequently leveraged survey questions to segment users. One question we asked was whether users were building for themselves or for a client. And we would use the answer to segment an account into an individual or business use case. However, with this line of questioning, we ended up incorrectly assigning freelancers who were experimenting with the tool but not yet professional users. Since then, we’ve refined segmentation surveys to use language our audience would say. We made sure that the options in the surveys are mutually exclusive, collectively exhaustive options. And finally, we’ve added questions with objective measures, such as the number of hours per week they spend on their online business. These changes have enabled us to better triangulate who users are, instead of purely relying on their own self-identification.

Another important aspect to keep in mind is that survey questions or profile information is often collected when you are setting up a user for the first time but can fail to evolve as users grow with your product. Webflow users who start out building for an independent use case often transition into freelance work as they get familiar with Webflow. Similarly, an Airbnb host may start out as an individual user listing their vacation home and really enjoy earning passive income in this way. Many hosts then grow with the product, acquiring and listing more properties over time.

To better define your segments, it’s important to supplement them with objective information, including in-product behaviors and third-party data, update your segment assignments regularly, and improve the way you ask survey questions where you leverage them.

At Airbnb, we could ask hosts a series of questions in their profile setup to understand if they fit in an individual or professional segment. However, other in-product behaviors were stronger, more dynamic signals, such as listing more than 10 properties, linking to a website, and uploading logos as part of the property listing’s photos.

At Webflow, we triangulate between in-product behaviors (e.g., site transfers, custom code), third-party data sources like Clearbit, and survey information to define segments for each audience.

For additional guidance and examples, Amplitude offers a behavioral cohorting worksheet to walk you through segmenting users based on behaviors or actions they take in your product.

Segmentation pitfall 3: Over-indexing on churned users

Finally, many product managers focus on the behavior of churned users. A churned user is someone who was actively using your product and went dormant for an extended period of time. Looking at trends about who churns from your product and why can be an interesting source of information, but can be limited in the impact it can have on improving retention for a few reasons.

First, churned users aren’t always a segment worth investing in. The probability that you can resurrect a churned user and get them back into a pattern of regularly engaging with your product can be low, and therefore, may not be worth the time or effort.

Second, churn reasons may not always lead to the right product insights. Although “churn reasons” from exit surveys or closing support conversations can give you context on what the user was thinking or experiencing at the end of their relationship with your product, they often do not give you the insight on how to change your product to reduce the likelihood that another user churns for a similar reason.

A churn reason that is often cited for most products is that users are not getting enough value out of the product. However, this information alone isn’t enough to determine how to reduce the likelihood that another user has a similar experience or reaction. In the case of Webflow: Are they not getting enough value because they found a cheaper solution? Did it take too long or was it too hard to learn how to use Webflow? Or, are they missing important tools or features that would make Webflow more valuable?

If they’re missing important functionality, then you might add new features to the roadmap. But if the tool was too hard to learn, adding features would make the problem worse. If they found a cheaper solution, the approach you might take to improve retention is adjusting the pricing model.

To avoid this pitfall, I recommend that product managers look further up-stream than churn events or reasons. Insights about what is different between users that activate versus those that do not, or those that engage regularly versus inconsistently, will be much more actionable for improving retention. And if you do need to leverage churn behaviors, recognize that this information is surface level and dig into the deeper why before drawing any final conclusions.

User segmentation can unlock lovability across many different types of users

To learn more about the various ways you can segment users and to see more real-life company examples, check out Amplitude’s Mastering Retention Playbook.


And if you enjoyed JZ’s perspective, be sure to check out her great courses on Reforge. To learn more about product management skill sets and career development, check out her Product Management Foundations course or Managing Your PM Career course.

About the Author
Headshot of Jiaona Zhang (JZ)
Jiaona Zhang (JZ)
SVP of Product at Webflow
JZ is the SVP of Product at Webflow, where she leads the product management, design, research, product analytics, product operations, and partnerships teams. She has been teaching product management at Stanford for the past 6 years and is the co-creator of the Reforge Product Management Foundations and Managing your PM Career courses. Over her career, she has coached hundreds of PMs on their careers. She has transitioned people from a diverse set of backgrounds (from engineering, design, sales, support, marketing, and operations) into the PM role, given them the tools to be successful PMs, and grown their careers into managers themselves. JZ is passionate about helping people articulate bold goals and then make them a reality.

More Best Practices
Image of Dillon Forest
Dillon Forest
CTO and co-founder, RankScience
Julia Dillon Headshot
Julia Dillon
Senior Product Marketing Manager, Amplitude
Image of Pragnya Paramita
Pragnya Paramita
Group Product Marketing Manager, Amplitude