Urban Farming

Product designer and user research

K+S wanted to deploy self-sustaining urban farming containers across Berlin. Before placing a single one, they needed to know exactly where to put them. The answer was a consumer app — a sustainability-focused map that would build a community, gather market data, and prove demand.

Team

Melissa Ma

Engineering

2platforms shipped — iOS App Store and Google Play

6custom map categories, each with unique icon, colour, and pattern

1city live at launch — Berlin, with architecture to scale

MVPproof of concept for a larger agricultural hardware initiative

The app wasn't the product. The data was.

K+S had a long-term vision: deploy self-sustaining, automated food-growing containers in urban environments across Germany. The technology was being developed in parallel by an engineering team. But before placing a single container, they faced a fundamental question — where, and for whom?

The brief that came to design wasn't "build an app." It was: build something that attracts the right audience, earns their trust, and generates the geographic and behavioural data needed to make smart hardware placement decisions. The app was a means to an end — a community-building vehicle and a market research tool disguised as a consumer product.

The target user was environmentally minded Berliners — people who already sought out organic produce, supported local businesses, and were motivated by sustainability. Understanding where they concentrated in the city was the whole point.

We weren't just designing an app. We were designing the proof of concept that would justify a much larger physical infrastructure investment.

Five things that made this harder than a standard app build

01 — Hardware placement depended on app performance. The app needed to demonstrate genuine demand and geographic clustering before K+S would commit to container deployment.

02 — Community and business buy-in had to be earned. The map only worked if local restaurants, urban farms, farmers markets, and stores agreed to be listed — before any user base existed.

03 — Scoped for iteration, not completion. This was an MVP with explicitly planned future phases. Every design decision needed to accommodate features that hadn't been scoped yet.

04 — Understanding a niche user through research. Bio shoppers and sustainability-motivated users have specific motivations and habits that differ meaningfully from general consumers.

05 — Marketing was off the table. Growth had to come from the product itself — word of mouth, discoverability, and genuine utility for its target audience.

Research changed the direction before design began

The initial brief leaned toward a novelty app angle — something eye-catching and marketing-forward. Research pushed back against this. The budget didn't support a marketing-led launch, and more importantly, the primary goal wasn't consumer acquisition for its own sake. It was data collection.

The research-led conclusion was a business finder map with a sustainability filter. This approach served every goal simultaneously: it gave environmentally minded users genuine utility, it created a reason for local businesses to participate, it generated the geographic data K+S needed, and it positioned the app as part of a community rather than a product looking for one.

The strategic logic: By slowly building a community and understanding where in Berlin sustainability-minded users were concentrated, K+S would be able to launch their farming containers exactly where they'd get the strongest reception. The app was the research instrument.

01 — Product vision. A project plan was established around a single north star: build the evidence base needed to deploy a self-sustaining urban farming container.

02 — Value proposition definition. Marketing-led approaches were deprioritised due to budget constraints and a mismatch with the data-gathering objective.

03 — Product research. Research into the target group reframed the product direction. A community map built around local and sustainable businesses emerged as the strongest fit.

04 — User analysis and ideation. Priority was given to farms, urban farms, farmers markets, and local stores — the destinations sustainability-focused users already sought out.

A map built on a system, not just a style

The core of the app is a location map filtered by business type. For this to work as both a consumer product and a data-collection tool, the categorisation system needed to be rigorous — clear enough for users to navigate, structured enough to generate clean, filterable data for K+S.

Six categories were defined, each with its own custom icon, colour, and repeating pattern that carries through to individual business profile pages. Any business fitting a category could be added — the system was designed to be community-fillable, not editorially curated.

The Urban Farming app — a sustainability-focused map of locally sourced businesses and events across Berlin.

Urban Farms

Growing produce within city limits — the anchor category tied to the container hardware vision

Farms

Traditional farms supplying Berlin with local produce, prioritised for their sustainability credentials

Farmers Markets

Regular and pop-up markets; time-based visibility tied to the events system

Local Stores

Independent retailers stocking locally or sustainably sourced products

Restaurants

Venues that prioritise seasonal, local, or organic menus

Events

Added late at client request — required extra thought on map labelling and hierarchy

Visual language — warm, grounded, and organic. Distinct from generic map apps without feeling niche.

The category system in practice — each type has a distinct visual identity across the map, list, and business profile views.

The details that made the system work

Smart recommendations without LLMs. The app included a "magic" button that surfaced the most popular or best-rated locations near the user. Built on popularity and rating signals rather than any language model, it delivered personalisation at zero infrastructure cost — predating the current wave of AI-powered recommendations by several years.

The recommendation system — popularity and rating signals surfacing the most relevant locations without LLM infrastructure.

City-specific onboarding. The sign-up screen adapted to the user's city. At launch, Berlin was the only supported market — but the architecture was built to scale. The onboarding communicated that scoping honestly, setting expectations without making the product feel limited.

Custom iconography and business profiles. Each category has a custom icon — the urban farm icon, a building with a small sprout on top, is the most distinctive. Every business and event has its own profile page using its category's colour and iconography as a visual foundation, giving the app coherence at scale without requiring editorial effort per listing.

Custom iconography for each category — the visual system that makes the map filterable and the data structured.

What worked, what didn't, and what this taught me

Research should redirect briefs, not just validate them. The initial direction was marketing-forward. Research showed it was the wrong frame for the goal. Pushing back early — before design work began — saved time and produced a more coherent product.

A category system is a data architecture decision, not just a UX one. The six categories weren't just navigation — they were the schema for the data K+S needed. Designing them well meant the app could serve both users and the client's research goals simultaneously.

Late feature additions need their own design pass. The events category caused a genuine semantic problem: events and places needed different labelling on the map and different hierarchy in the list view. It worked, but it was a seam that showed.

MVPs built for scale need honest constraints communicated to users. City-specific onboarding set expectations without apologising for the scope. Users who understand the constraint often become advocates for expansion.

More work

cara care

Cara Care was a technically capable app sitting inside a fragmented, brandless shell. The challenge was to build a visual identity, a design system, and a data-driven tracking experience rigorous enough to earn DiGA medical device approval in Germany — without losing the human warmth that keeps patients coming back.

View Case Study →

Qlair

An app aimed towards technicians, who need to keep track of air quality and filtering systems across huge buildings.

View case study →