Their plants were fine. It was the apps they gave up on.
The people I spoke to loved their plants. What they didn't love was an app that only ever nagged, never provided enough information about what to do, and turned one missed day into a red “overdue”. It gave them nothing to come back for, so they stopped opening it. I wanted it to feel more like a friend who knows plants: someone who tells you what to do today, and is happy to see you when you come back.
Reminders are useless if they're not helping you to achieve your goal. People did not want to wonder around by asking questions to the app to achieve their goals, it's tiring and not worth to the effort.
I worked on this alone around 4 months this year (2026), from first time drawing things on paper to testing the prototype. Working solo meant testing with the same 10 people I interviewed, all found through three Facebook groups. They gave me their time twice, and the test results are theirs.
I read 200+ low-rated Google Play reviews of the ten most-downloaded plant apps, by hand and cross-checked with three AI tools. I spent time in three Facebook plant groups, read Sapolsky’s Behave to understand how reward works, then interviewed and tested with 10 plant owners. Every design decision had to point back to one of those findings.
With the new task card, all 10 people knew what to do without any help. Only 3 managed that with the old one. They also worked it out twice as fast: about 3 seconds instead of 6. It’s still a concept and hasn’t shipped. I would prefer to make the test with MORE people. The next test checks the thing that made people leave: whether they come back the day after a missed task and goals.
Sneak peek at what’s coming in this project
No story here, only a look at two animated flows. The problem and the reasoning start below, with screens alongside.
Plant apps left people unsure, and one missed day made them quit
Three sources, each answering a different question.
Losing a reward you expected feels worse than never having it. A reset streak is exactly that.
I associate the problem as Dominoes, everything in the apps made users closer to quit, not just one domino.
“I’d skipped a task to give my plant time to recover, and the app yelled at me in red. Deleting it took no effort.”
The brief changed from “remind people to water” to three goals.
Four screens, and the question each one answers
None of these started as a feature idea. If a screen couldn't name the question it answered, it didn't make it into my design process
The app in one picture · four main tabs
The friends’ stories row: testers said it pulled them away from their own plants. The account wall: reviews kept naming “complicated menus and logins”. Recommended plants: one review called a rival app “more a plant shopping app”.
Three task cards became one: what, how much, when
Home, Track and Rooms each had their own task card, so one task looked three different ways and people had to learn each one. I drew 7 drafts chasing one card that’s accessible, quick to scan and clear about what to do. Early drafts hid the actions or left out the amount. I tested the best two against each other, and the winner now runs on all three pages.
7 designs in a week · 4 shown
B and A are the two I tested. I expected B to hold up because it looked clean and calm. Without an amount, nobody could say how much water to give.
The test · A vs B
I showed 10 potential plant users both cards, one at a time, and asked: “What does this card want you to do?” Half saw A first, half saw B first. No hints.
The same ten people I'd interviewed, so they already knew the problem. It tells me A is clearer than B. Whether it works for everyone needs a bigger group. The 3 of 10 who still mixed up skip and delay sent me to the wording problem, its on below.
How the app gets to ¼ cup
Plant care guides say to water up to ⅓ to ¼ of the pot’s volume. I used ¼ as the safe start, since overwatering is the number one killer of indoor plants. The app does the maths, which beats “water well” or a number someone picked for their own plant.
How to tell and understand the difference between "Delay" and "Skip" actions.
How I resolved it: I kept both names and added one line as helper text “Skip just this time”: “Keep the schedule. Next watering 3 Jun.” Now its more clear to understand the difference before you tap.
Version A task card went into all three pages as the result of the test, users do not have to adapt or learn new task card system, which was a bad idea for creating consistency across the app and most importantly, card answers every question about completing a task.
A tester broke my goal system without trying to
During a usability test, a tester typed 400 trees into a 500-tree goal and nothing stopped her to ruin the goal progress. I only realized it by watching it at the same time and it dropped like a bomb. It looks and sounds very easy to avoid such problem, got my lesson from it.
The goal, before and after the fix
Anyone could type any number
I assumed a few bad numbers would get lost among hundreds of honest ones. One phone at home can finish or wreck the goal for everyone on it.
Count the saplings, not the claims
A phone can’t prove a tree went in the ground, but saplings arrive in a crate with a delivery slip. Your count comes from the lead’s tally, so 400 isn’t a number the screen accepts.
Four ideas I brainstormed first
Each one had the same flaw: strangers doing unpaid checking while an honest person waited.
The fix worked, but now the page was full of numbers
A newcomer saw a spreadsheet before they saw a Join button. So I moved the counting to its own page and kept the goal page simple.
A static screen can't be abused, so this problem appeared ONLY because someone used the prototype. Each fix went back into testing for the next person to try to break.
Personal outcome was to explore a system with every scenario possible whether happy, sad, frustrating ending.
Still open: Flow doesn't include the perspective of the HOST, this would be my next move to complete the flow before testing with many users. The crate count itself hasn’t been re-tested yet either.
A missed day costs nothing, and the streak only goes up if user continues
A streak resetting to 0 is the third row. From Behave, Sapolsky.
Getting less than you expected feels worse than getting nothing. A lost streak is exactly that.
One missed day, two ways
What I kept, rebuilt and cut
Streaks pause, never reset
The count waits for you.
The card nudges the worthiness
“Today makes 6” sits above the task.
No ranking against strangers
Only inside a shared goal, off by default.
No surprise rewards
A surprise feels great once, then you start expecting it. When it doesn’t come, it feels like a loss.
No countdown to the hour
For a plant in a pot the day matters and the hour doesn't, so the card says Today, Tomorrow or a date. No rushing or causing anxiety about timing.
Only show what the user actually did
- 1What did I do overall? “47 waterings, last 8 days ago” is a fact you can check.
- 2What did I miss? A gap is named in plain words, never graded the user.
- 3Tabs gone. One page to scroll, so nothing hides behind a tap.
It falls when you prune, it's a guess on a page of facts, and checking it means measuring the plant anyway..
If I can't answer “how do you know that”, it doesn't go on the screen.
Nothing can be taken off you. Missing a day costs nothing, and every number on the plant page is something you logged yourself.
Priorities! Today’s task replaced friends’ stories, and sign-up moved after the first care plan
A stories row and a sign-up flow. Both were finished, both looked good, both answered the wrong question first.
Home gave its best strip to other people
In testing, people told me the stories row pulled them away from their own plants. When you open a plant app you want one thing: what does my plant need today? The social idea wasn’t wrong, some owners loved seeing other people’s plants, but Home was the wrong place for it.
Stories were fun, but they pulled your eyes away from your own plants. They’re also hard to keep kind and easy to misuse, so they didn’t really belong to the app. I redirected this "social" problem by designing "Ask your friend" card, you will see 2 container below.
Recommendations are just more stuff to scroll through, and nobody opens Home hoping for more stuff. The toolkit puts the things you actually reach for right there instead.
The account ask comes last, after the app has been useful
People open a plant app when something has already gone wrong, which is a bad moment to ask for an email. In the reviews, nearly half of the complaints were about paywalls, surprise charges and “complicated menus and logins”.
The business needs accounts. The user needs help first.
So the app helps you first and only asks for an account once you want to stay, which is why first experience is very crucial for business and its' users.
Everything I promised is free. You only pay once your collection grows.
Why three? It’s enough to see the app can easily handle more, it gives you a reason to start your own collection, and it’s still generous. The upgrade prompt only shows up after your 3rd plant, so the business earns from people who already stuck around.
- Your own care history
- Help with a sick plant
- Double points for paying
Nearly half the bad reviews were about surprise charges. So the price is never a surprise.
Home answers “what does my plant need?” before “what are my friends doing?”, and the app proves it’s useful before it asks for anything back.
A warning can never look like a care tip, and late is written gently
Hand-off: the Figma file carries sizes and tokens, the working prototypes carry behaviour, and every task-card state has its own words, never colour alone.
One green was doing four jobs
“Add fertiliser” sat next to “toxic to your cat”. Now a safety warning can never look like a care tip, which is why late is written in words and feeding is brown.
Every option tells you what it’ll do
Tap ⋮ and the same sheet opens every time. Each option has a short line underneath, so you know what happens before you tap.
Set up in the first minute
Every answer changes how the app behaves, and each one can be tried before it is accepted.
Six sizes, four steps, one typeface
Where it leaks: screens drawn before the library still carry off-scale values. Auditing them back is on the list.
One two-week test would tell me if people stay
Nothing here has shipped, so there’s no real usage data yet. People left other apps after one missed day, so this first study is built around coming back: the same people use the app for two weeks and skip one day on purpose.
Planned, not run yet
The task card is one reason people leave, not the only one. Each checkpoint tests a different moment where someone could quietly give up.
Time from install to the first plant added and the first task on Track.
Pass if most people get there in under 3 minutes without asking for help.
Count how many people mute or turn off notifications, and ask how the last one felt.
Pass if fewer than 1 in 5 mute them, and nobody calls them annoying.
Everyone skips a day on purpose, then opens the app the morning after.
Pass if they come back, and nobody says “guilty” or “told off”.
Ask about a plant that looks unwell. Watch whether they open the app or ignore it.
Pass if they turn to the app for advice instead of feeling judged by it.
Ask the Day 1 confidence question again, and “would you keep this installed?”
Pass if confidence goes up and most say yes without thinking about it.
What this project taught me
I didn’t expect YourNature to be this hard, or to take this much of my time. Career coaches say to pick a project that stretches you but stays within your level. At the start, this one was beyond mine.
Only after finishing the designs did I see how much it had taught me. It took far more hours than I planned, and I enjoyed them anyway.
ch 06
Let’s work together
I’m looking for a product design role where I can do both the research and the craft. If that sounds like your team, I’d love to talk.
