Ambie 3.0 was difficult to build, not because of code complexity, but because of the brutality I had to learn as I slashed features and rebuilt entire designs.
Original Plan
The original plan for Ambie 3.0 was straightforward: rebuild the UI to support Windows 11’s new beautiful material called Mica. Previous versions of Ambie was built with Windows 10’s Acrylic material, and while it was nice at the time, Mica has proven to be an amazing successor. Ambie had to have it, and it wasn’t going to be difficult to add. Or so I thought.
Original Plan, Take Two: Now with more data
Simultaneous to experimenting with Mica on Ambie, I’ve been obsessing over the anonymous telemetry that Ambie was collecting. Everyday, I checked AppCenter for any patterns. I was trying to understand how people were using Ambie, and if there were any ways I can improve the app based on their behaviour. Here were things I noticed:
- Ambie was primarily used on weekdays
- The sleep timer feature (used to auto-pause playback and allow the PC to sleep) was primarily used for 60 minutes, not the usual 15 minutes that is common when going to sleep.
- Some of the most common feedback was adding longer sleep times, such as 2 hours or 4 hours.
It was clear that people were using Ambie for productivity work on their PCs. It sounds obvious now, but since I built Ambie to help me relax and destress, it took some time for me to overcome my biases. I eventually accepted that Ambie needed a full featured, built-in Focus Timer.
And thus the original plan changed after understanding the data and signals from customers. I needed Mica and a Focus Timer feature in Ambie 3.0.
Initial designs thrown out after testing
As part of my usual development process, I sketch everything out by hand and try to see if it works. Often times, I realize if a design on paper doesn’t look great so I can iterate before writing code. But for Ambie 3.0’s initial designs, it was difficult to understand if the design would work effectively. I needed a prototype.
So, the design below did actually go into a prototype. They went as far as an alpha build deployed secretly to the Microsoft Store. But after testing it thoroughly, it took me a week to accept that the layout did not work on a Windows app. The primary issue was it didn’t make effective use of the horizontal space. As a user, I got frustrated that I had a lot of empty space that was wasted. I knew that if I released Ambie 3.0 at this state, it would not be good enough. So I scrapped it a week before I was scheduled to release. Roughly two weeks of prototyping and one week of testing thrown out.
Initial features, also thrown out
Over the course of design and development, numerous “nice-to-have” features were written down. I thought that many of these would help Ambie’s focus timer shine even more. But the reality was that I didn’t have the time to build them and they were superfluous. Not having them there did not hinder Ambie 3.0’s potential, and I knew I could always add them later if I wanted to. Here were some of the features that were cut:
- Motivational text to encourage people as they focused
- Smarter notifications that gave people a warning when focusing was about to start or end
- More design/customization options for the timer UI
- Achievements when completing a focus session
- And many more
Was it worth it?
Recently, I spoke with a mentor and learned of the Project Management Triangle. For each project, you have to balance between scope, time, and cost. The quality of your product will depend on how you balance those three.
For Ambie 3.0, I inadvertently drove the product quality up. I slashed the scope and reduced cost by cutting features, and I delayed the launch by an extra month.
The result was amazing
- 110% increase in monthly active users
- 70% increase in subscribers
- Rose to #1 white noise app in the global Microsoft Store
- Named one of 2022’s best new productivity apps by FastCompany
Learnings
- When it comes to landing a product, slashing features is just as important as building features to drive up product quality. It gives you time to really focus on the important features.
- If the UX is not good enough, don’t ship it. Build it again, or lessen the scope so you can build a better UX. I learned from someone that “it’s better to deliver less, than to deliver worse.”
- Use data effectively. Ambie’s Focus Timer feature was based on months of data collection and analysis. However, it’s important to balance numbers with real user feedback as well. Use data and feedback to understand user behaviour first, and then solve those problems and meet your customers’ needs. Don’t think of a feature and find data to confirm your biases. You’ll almost always find data that can be interpreted to meet a requirement.
