To Self: Where Writing Meets Analytics
Overview
ToSelf is an iOS journaling app that turns writing into data. Users write private messages to themselves, and the app surfaces keywords, sentiment, and emotions through analytics. The goal is to help people notice patterns over time, without a chatbot or an algorithm telling them what to think.
At a glance
Role: Lead Product & UX/UI
Designer Scope: UX strategy, information architecture, interface design, brand identity, web design
Timeline: 2024 to 2025
Background
I joined ToSelf in 2024, taking over design from a previous designer. The app already had a working product, but design work across the board needed attention: the underlying structure, the visual design, and the brand itself hadn't been fully resolved.
One clear example was how messages were organized. Every message a user wrote had to live inside a notebook, one of eight fixed categories like Work, Family, or Health & Fitness, and inside that notebook, messages were grouped further into threads. It worked, technically, but it asked a lot of a user before they had even started writing.
Challenges
-
Inheriting a project is always a challenge. It's not a blank page, and there's usually some resistance to change from a team already invested in what exists.
-
The information architecture itself: how notebooks and threads organized every message a user wrote.
-
A lack of brand identity. This wasn't product-related, but I also had to take on branding and help define what the brand looked and sounded like; Logo redesign was also part of the brand identity process.
Project Scope
-
Redesign the core organizational model for messages
-
Full UI and visual refresh (color system, iconography, hierarchy)
-
Brand identity: logo, brand voice, guidelines
-
Web design aligned to the new brand
-
Feature work: image-on-message, premium purchase flow, settings and message detail redesigns
UX Process: Organization 2.0
Research
Early in the redesign, we ran beta testing with around 10 users, recorded live while they downloaded and used a TestFlight build for the first time. This wasn't a moderated lab study. It was closer to watching someone try the app cold, the way a real first-time user would.
Feedback on the app overall was mixed, but one pattern was consistent and hard to miss: people found the two-level organization confusing. Having to choose a notebook, then choose or create a thread inside it, before they could finish writing a single message, added a decision cost to something that was supposed to feel immediate.
Even before saving a new message (formerly known as a note), the user had to go through a two-layer organization process. This was something that was flagged by the majority of users during the beta test.
Simplifying the Structure
Removing an entire layer of the product is a harder sell than simplifying it. The notebooks model wasn't broken by any technical measure. It worked. It shipped. People were able to use it.
Still, I thought this two-layer flow could be simplified into one. Users were confused about the difference between notebooks and threads, and from a user's standpoint, I understood that confusion. At the same time, I understood what the CEO was going for: a very organized, meticulous way to keep messages, formerly known as notes.
But two nested decisions stood between a user opening the app and actually writing. That's real friction, especially in a product whose entire value proposition is removing friction from reflection.
Overcomplicating is always the enemy of good design. To test my hypothesis, I started wireframing a simpler flow built around threads alone. A lot of engineering work had already gone into building the existing two-layer system, so I started by pitching something smaller: let users organize their thoughts using either notebooks or threads, instead of requiring both.
After iterating and getting more feedback from other stakeholders, I presented Organization 2.0: a new flow that eliminates notebooks. New messages are saved in threads only. That made saving a message easier in two ways: threads were no longer required, so people could assign one later if they wanted to, and there was no longer a fixed set of categories to choose from first. A message is now either uncategorized or assigned to a single thread, with no nesting. Threads are user-named and user-tagged with an emoji, sitting at one flat level in the Library.
A few of the early explorations for the Organization 2.0 flow. Trying different structures helped me land on something simpler and easier for users to move through.
The new flow is more straightforward. Users can assign a thread right from the Create New Message screen, or skip it and do it later.
Organziation 2.0 - Impact
In June 2025, once this shipped, we ran a second usability test: 15 testers, over 10 days, remote and unmoderated through BetaTesting.com.
The simplified structure held up. Alla, one of the testers, put it plainly: "Moving entries into threads is easy. The app generally is very easy to navigate. The interface is intuitive." Josue went further, saying the app was already so easy to navigate that built-in shortcuts to threads and the library barely felt necessary anymore.
The overall System Usability Scale score came in at 88, well above the 68 considered a good result, with a Net Promoter Score of +13.
Visual & UI Redesign
Beyond the organizational redesign, I led a visual refresh of the product:
-
Color system: moved to a Tailwind-based palette (blue/500 as the primary brand color, with a restrained orange accent), replacing analytics chart colors that were too loud for a product meant to feel calm and non-judgmental.
-
Iconography: introduced Phosphor as a single, consistent icon system, replacing a mixed set that undercut the product's sense of polish.
-
Hierarchy and minimalism: simplified spacing, type scale, and visual weight across core screens for more clarity and less clutter.
Branding
Redesigning the organizational model surfaced a bigger question: what does ToSelf actually sound and feel like? I led the brand work that answered it.
-
Logo redesign: a new handwritten-style wordmark that better matched the product's calm, personal tone.
-
Brand voice: reintroduced Stephanie Schwab's brand voice framework to define ToSelf's character, tone, language, and purpose. Friendly but never patronizing, direct but never clinical, warm but never overpromising.
-
Brand guidelines: documented logo usage, color system, typography, iconography, and copy rules (sentence case everywhere, no exclamation marks, never interpreting the user's emotions for them) so the voice would hold consistently across product and marketing.
-
Web design: extended the same brand system into the marketing website, drafting key messaging directly from the brand voice work rather than writing copy in isolation.
The Full Scope of the Work
Redesigning the structure and the brand didn't cover everything I worked on. I also touched several other parts of the product that people interact with every day. Some of the features that I've worked on are:
-
Image-on-message: attaching photos to journal entries
-
Premium purchase flow: a clear path from the free version to a paid one, supporting the app's freemium model
-
Settings redesign
-
Message detail view redesign
Key Takeaways
1. The highest-impact design decision is sometimes removing something, not adding to it. Notebooks weren't a bad feature in isolation. They were an extra decision standing between a user and the thing they actually came to do. Simplicity meant subtraction, not iteration.
2. Small-sample research, argued well, can move a stakeholder more than a large dataset argued poorly. Ten recorded beta sessions were enough to build a case for a structural change, because the pattern was consistent and the reasoning behind it was documented, not just asserted.
3. Inheriting a product is a different design problem than starting one. Diagnosing what to keep, what to fix, and what to remove from someone else's decisions required a different kind of judgment than greenfield design. It meant separating what genuinely didn't work from what simply wasn't how I would have built it.