01 — The Problem

New to gardening.
Short on time, space, and energy.

Like most people who start a garden, I came in with enthusiasm and left the first season confused. What went wrong? What should have gone next to what? When was I supposed to start those seeds inside? I looked for an app that could answer these questions and found nothing that actually worked — tools that were either too simple to be useful or too complex to bother with.

Then I discovered two methodologies that changed how I thought about the problem: companion planting (which plants support or suppress each other) and square foot gardening (how to maximize yield in a small space through density and planning). Together they formed a system — and systems, I know how to design.

The gap wasn't data — there's plenty of gardening data online. The gap was decision support: given my specific space, my growing zone, my time constraints, and what I want to eat — what should I plant, where should I put it, when do I need to act, and what do I do when things go wrong?

No existing app answered that question. So I built one.

02 — The Systems Challenge

One plant, thirteen states.
Two user realities. Infinite timing variables.

The hardest design problem wasn't the visual layout. It was the underlying state model — what it means for a planting to exist, progress, fail, or succeed, and how that maps to a user's actual experience in the garden.

The 13-state lifecycle I designed maps a planting instance from first intention through final harvest or failure — accounting for every branching path a real plant takes:

The companion planting logic, which initially seemed like the complex part, turned out to be more elegant once I realized it operates at the plant family level — alliums, asteraceae, solanaceae — rather than requiring custom rules for every individual plant pair. That insight collapsed what could have been thousands of relationship rules into a manageable taxonomy.

Two views, three tabs

The planning interface needed to serve two mental models simultaneously: spatial thinkers who plan by grid and sequential thinkers who plan by timing. I designed both — a visual grid view and a list view sortable by timing or alphabetically — unified by the same three-tab structure: Active (in the ground now), Planned (placed but not yet started), and Done (completed, failed, or saved).

The three-tab structure did something important beyond organization: it defined the seed inventory. A planting only counts toward seeds needed once it's in the Planned tab — giving users a clear, accurate picture of what to buy before the season starts.

03 — The Build

From personal tool to
App Store product.

GrowWise began as a tool I was building for myself in Lovable.ai — a prototype that worked well enough to use in my own garden. The moment it became a real product was when I realized the underlying system I'd built had value for anyone who gardened the same way I was trying to learn.

04 — Testing & Iteration

Every weekend in the garden
was a usability test.

Testing was 20 beta testers — roughly 10 on iOS, 10 on Android, all friends and family — and me, using the app every time I worked in my garden. That second part mattered most.

The iteration cycle was simple and direct: spend Saturday in the garden using the app, notice what was confusing or missing, build the fix, ship Sunday. Repeat. Over two growing seasons this produced a cadence of continuous real-world usability testing — not in a lab, not with a script, but under actual conditions with actual stakes.

Key features that emerged from this process: the timing relative to last frost (I needed to know how many weeks I had, not a calendar date that meant nothing to me), the seed quantity calculator (I kept buying too many or too few), and the harvest guidance per plant (I kept harvesting too early or missing the window entirely).

05 — Reflection

What I'd do differently.