Skip to main content
Tech Talk

User research to increase retention rate (Part 2)

By December 17, 2019No Comments
Amazing illustration by Zamax

Collecting, brainstorming, implementing & results. 

It’s been a while since I had been able to defeat the Evilish Resistance inside me (the War of Art – Steven Pressfield) to sit down and write the second-and-last part of this series. 

What I have learned from working with g an SEO specialist is that it is much easier to write small, niched posts then move up to big ones than the other way round. When we finally come to the last post, our energy and enthusiasm for the topic must have died down significantly, yet we’ve accumulated a lot of knowledge about the matter, making writing it a piece of cake. 

A brief summary of the previous part

I started with my ceaseless urge to write a topic about user research, how I chose the tools and user segments. Then, I spent a majority of my space explaining how I built my survey/interview questions, and all the valuable lessons I’ve learned along the way.

Feedback for the previous post

I received feedback from my girlfriend to:

  • make the article title clearer (yeah, I didn’t know what to expect if thrown in the face “user research to increase retention”). In retrospect, I had a fancier name for it, yet I chose to go for the boring and useless one. 
  • refer to the data of free users’ retention rate instead of both. This is legitimate since retention rates between free users and premium users differ largely from one another. 

Let’s see if I can improve them in this post.

Collecting data

After collecting and processing the data, I was in awe with the insights. Here is the list of purposes people use Habitify for,  with the top having the most respondents. 

  1. Track the current progress of my habits
  2. Remind me so I don’t forget my habits
  3. Form new habits
  4. Improve my life in general
  5. Motivate me to stay on track every day

What’s funny is that Habitify was designed to help people with (3) forming new habits, but it was merely used to (1) track the current progress of their already existing habits and even recurring tasks. 

Both me and Peter (my founder) were struck. 

People came to us when they already had an idea of what to do. They just couldn’t stay on track. 


“The” Process

So I’ve identified the purpose. It’s time to brainstorm the solutions. My first step when brainstorming solutions is to always identify the users’ real problem. 

I want to contribute this mindset to my boss – Peter – and also to Maslow’s Law of Instrument. While Peter has been teaching me to think of designing product problem-wise instead of solution-wise, Maslow (who I believe is famously known for the Hierarchy Need) further solidifies the importance of problem-solving:

If all you have is a hammer, everything looks like a nail. “

I’d write another post for this subject, but right now, this article from Zhang – Head of Product at WeWork, suffices: Building a Minimum Viable Product is Like Serving Burnt Pizza

So rather than diving straight into a solution, my typical process often looks like this:

  1. Identify the real problem(s)
  2. Check how many percents the current product can solve this (instead of outright disregarding all the useful functions – which is a foolish action – take a look at how much they can solve the problem and what’s left open for improvement)
  3. Devise a solution to fill in the blank or replace the whole current solution. 

In reflection, my Dream Internship at P&G taught me the very same thing. This process of thinking has helped me tremendously, as I will not miss out on the big picture (user problems) while I can devise innovative solutions that do not overlap with current ones. 


Back to Habitify. The top two purposes people use Habitify for is to track and to remind, so, what are their problems?

When I looked deeper into all the feedback emails and hours of online call notes, I realize something very interesting: Contrary to our thoughts that habits must be repeated at a high frequency (at least 5-6 times a week, if not every day), people have varied perceptions about habits. 

  • For some, habits are recurring little household tasks like washing their mattress once a month or cleaning the floor every 2 weeks (come on folks! how could we think that someone would need to be reminded to wash mattress by an app? Should we make a “Mattressify” app or something?) 
  • For others, habits only need to be done a certain number of times per week, say, read a book 3 times a week, or go for a run 4 times a week. 
  • For others of others, habits must be done strictly every few days. A very funny example that belongs to my user in one of their (endless) request email is washing her hair. 

“I want to wash my hair every 2 days, but Habitify failed to help me track that. If I set up a habit to repeat every 2 days, Habitify will follow that without any flexibility. Say, if I start it on Monday, Habitify will repeat it on Wednesday. But if I forget to do it on Monday (you forget to wash your hair?) then I do it on Tuesday, Habitify, then, still repeats the habit on Wednesday rather than supposedly Thursday.”

These footprints slowly lead me to the real problem:

People have very diverse habits that need to be tracked differently, and no current tool can do that efficiently within one interface

When I finally came to that declaration, another “aha” came to me. Now, I realized that todo-list apps have never been my competitors (sorry Todoist!).  People use a habit tracker as a complement to their todo-list to declutter the already-busy interface stuffed with nested tags and folders. This is how I came up with a special landing page addressing the need. 

Lesson #1: Starting with problems gives me the big picture

Since people have different perceptions of habits, the needs to be reminded vary accordingly.

  • For people who want to repeat their habits daily, a timely reminder is enough (which we did outstandingly).
  • But for those who are seeking a touch of motivation, they want more human and personalized reminders.
  • For those who are procrastinating, similarly, aggressive nudging works the best. This is where we fall short. 

And there we go, the real problem:

People want to customize and personalize how they are reminded to do their habits, which current tools fail to do so.

Lesson #2: People always want to be in control, and they will leave the product if they can't. Customization and personalization, minor as they might sound, can contribute a lot to bringing them back. 

Implementing the solutions

For people that have different tracking needs, here are our solutions:

  • We provided more choices to repeat their habits. Instead of just specifying the days, we have every X days and X times per week. This directly addresses the need to track in different ways. The hardest part of this, which admittedly has been a puzzle to the team, is how to reflect the stats correctly. Since the frequencies of each habit are different, there’s no numerically and representatively correct way to aggregate them by one single number.  
The more flexible habit tracking is, the more difficult to calculate the statistics.
  • To partially make up for this incompetence, we allow people to dig deeper into their data instead of clumsily generalizing numbers. We added some more metrics to the Progress tab. Now, people can see their daily, weekly, monthly, yearly progress in both numbers and percentages. What’s more interesting is that they can actually compare habits with each other see which one is falling behind. 
Look! I spent 1/3 of this year to wake up early at 5:30 AM!
  • In the Journal screen, we replace the “reminder” information with the “streak” information”, allowing people to know how many times they have performed a habit. This also acts as a motivator for a lot of people at first glance when they open the app. 
Certainly more helpful than the reminder time!

For people with multiple reminder needs, here are our answers:

  • The first solution is to diversify the frequency options for habits, which I have mentioned earlier. 
  • Next, we allowed people to “snooze” from the reminder. This is a workaround for people who needs to be constantly nudged to do a habit.
  • We personalized the copy for the Daily Briefing and also the current reminder. Instead of a simple: “It’s time to Read Book” or “It’s time to check your morning habits”, we inform them the streaks they are on, when the last time they checked in was, what their first habit of the day is, how many habits they have today, and the likes (depending on how they are performing). If you’re new to microcopy writing, I’d suggest taking a look at this book by Kinneret Yifrah, it’s an amazing read with lots of helpful examples. 
  • We allowed people to choose when the default Daily Briefing appears. Previously, it appears at 7:00 AM and 9:00 PM, but many people wake up and hit their pillow earlier than that, which render the Daily Briefing useless in that sense.
A slider to change the time of day that the Daily Briefing will appear

Note: We’re building multiple check-ins solution to address the need of those who have repeating habits like drinking 8 cups of water per day

Admittedly, these are not permanent solutions, I can tell you that. Sometimes I have a second thought about what has been implemented. However, with the current time and manpower resources, it’s best that we could do at the moment. Adopting the Agile mindset, we never settle down on anything, but we deploy, monitor, and then improve constantly. 

That mindset relieves my mind. 


The immediate response after the release(s) was people’s appreciation of our listening to their feedback. I will not directly attribute the five stars and lots of positive tweets to this release alone – that could be single-minded – but I know our branding as a good-listener-app-makers has been greatly improved. 

After a quarter of implementation, we’re glad to see the effectiveness of our problem-solving process. The retention rate for all users has increased 10%, meaning that out of 1000 new sign-ups each day, I have “rescued” 100 more people!

Here’s the 30-day retention rate of all users in Nov 2018: 

Source: Amplitude

And here’s the 30-day retention rate of all users in November in 2019 (we changed our analytics tool)

Source: MoEngage

In Japan, one of my key markets, the retention rate jumped from around 32% to 44%! 

Final words

There are three things worth mentioning here:

Why I don’t use Free users’ stats? 

The previous tool (Amplitude) allows me to split the retention rate based on user attributes like Premium status, but the new one (MoEngage) does not allow me to. Therefore, I cannot compare based on the user’s attributes. Personally, I don’t think the new features make much of a difference to Premium users’ retention rate since they have already paid for it. But rationally, it’s still open to deeper investigation. 

Why only 30-day retention?

I adopted MoEngage just October this year, so I couldn’t compare the before-after 1-year retention like I really wanted. 

Did I use this with other methods?

Yes, in fact, the number you’re looking at is the compound effort of many other practices like improving the onboarding flow with emails and push notifications, building a community or overhauling the customer success workflow. I will cover these topics in other articles. 

The more I write about this, the more I realize how much I love working with digital products. Even seeing a tiny uplift of 1 percent in retention rate, or a thank-you-Habitify email from my customer is enough to make my day. And my day would be so much better if I could receive some feedback from you, who have patiently read till these last words. You know you have my appreciation. 

Leave a Reply