Product Design Case · Feb-May 2026

Compawnion

Pet adoption that fits your life.

A mobile app that helps first-time adopters find the right shelter animal, by honestly mapping daily life before browsing.

View interactive prototype →

Tools Figma · Miro · Claude

Survey n = 120

Interviews 4 qualitative

Testing 4 Remote Usability-Tests

Timeline February-May 2026

The Problem

The desire is there.
The decision feels overwhelming.

When people want to adopt a pet, they quickly face the same problem: too much information from too many sources, and the real question goes unanswered: Does this animal actually fit my life?

84%

considered adopting from a shelter.

>50%

feared rejection from the shelter, before making first contact.

44%

higher fear of making the wrong decision among first-time adopters vs. experienced ones.

“It feels like too big a task. I really feel like this task has overwhelmed me in my head.”

Interview, semi-experienced adopter

These aren’t information problems. They’re a missing tool for honest self-assessment, and a lack of confidence at the first step.

Two moments where adoption regularly breaks down.

Desire

Research

Choosing an Animal Critical Moment 1 Suitability for Daily Life

Initial Contact Critical Moment 2 Fear of Contact

Arrival

Process & Key Decisions

Four decisions that
shaped the product.

Profile-first instead of browse-first.

The classic flow shows animals first and leaves it to the user to figure out the fit. My research showed: the problem isn't the supply of animals, it's the lack of self-assessment. Compawnion flips the flow: users define their daily life before they browse. This reduces aspirational bias and makes matching more honest.

“Profiles saying ‘sweet, friendly, great’ don’t help me. What helps: likes to sleep on the couch, needs a moment in the morning, goes crazy when someone rings the doorbell.”

Interview, first-time adopter

“It would help to have access to good information that isn’t just buzzwords but can actually be matched against my daily life.”

Survey, first-time adopter

First-time adopters as the primary target.

Compawnion serves anyone looking to adopt — but I designed for the user where the problem is most acute: the first-time adopter who overthinks every step and abandons the process before making first contact. The hardest to serve, and the most underserved.

The Situations-Check as a contact anxiety intervention.

The critical moment isn't choosing the right animal — it's the transition from observer to applicant. A solution that prepares users for first contact in a structured way: what to expect, what the shelter expects, what they can say. When they hit 'Send,' they feel already decided.

“I always feared they would categorically reject me as a pet owner because of my living situation in a shared flat.”

Interview, first-time adopter

Matching quality depends on shelter data quality.

I decided to build the user side first — it's the part that can be validated, and a working user flow is the prerequisite for bringing shelters on board as partners. This isn't a forgotten problem. It's the next step.

Persona

Designing for Maja, the first-time adopter.

Compawnion serves anyone looking to adopt — but every design decision was made for one specific user: Maja. She represents the hardest case: analytically minded, risk-averse, informed — and still unable to take the first step. If Compawnion works for Maja, it works.

Behaviors & Habits

Risk-averse — weighs every option twice before acting.
Obstacles delay or even halt her decisions.

Pain Points & Frustrations

Fear of misjudging the animal’s daily life.
Overwhelmed by scattered or contradictory information.
Often doesn’t know what to make of the information.

Needs & Goals

A clear assessment of the fit.
Transparent and comparable animal profiles.
Practical advice for everyday life.
A stress-free and transparent process.

The Solution

Two moments where adoption regularly breaks down.
Two interventions.

Compawnion is a decision support tool. It helps (first-time) adopters discover animals that genuinely fit their daily life, rather than browsing and hoping for the best. The goal is not to overwhelm with options, but to build the structure and confidence users need to take the next step with clarity.

View interactive prototype →

Moment 01

Choosing the right animal, the Alltagscheck.

The Alltagscheck questionnaire captures the relevant dimensions of Maja's daily life. The answers feed directly into the matching logic: each animal in the list shows an Alltagspassung score. Tapping it reveals a breakdown by dimension, showing how well the animal's needs align with Maja's profile. The number is not a verdict. It's an invitation to look closer.

Moment 02

Making contact, the Situations-Check.

Before reaching the contact form, Maja sees a structured overview of how her profile compares to the shelter's requirements. A low score on individual criteria is not a dealbreaker — a tip reminds her that meeting in person often reveals more than any checklist. Once sent, a confirmation screen shows clear next steps, giving Maja visibility into what happens next and reducing the uncertainty that follows first contact.

Design

A visual system built around calm confidence.

The design target emotion for Compawnion is not excitement — it's calm confidence. Every visual decision reinforces that: a warm but unsentimental palette, rounded but structured type, and a matching system that communicates clearly without overwhelming.

Logo

The Compawnion wordmark uses Knicknack, a rounded display typeface. The brief was specific: carry the playfulness of the name — a portmanteau of 'companion' and 'paw' — without tipping into the generic visual language of pet product brands. Rounded forms signal warmth and accessibility, appropriate for a product that addresses anxiety and uncertainty.

Typography

The type system uses a single superfamily: Nunito for display and headings, Nunito Sans for UI text and body copy. A shared skeleton across both cuts maintains visual coherence without requiring a second typeface — hierarchy comes from weight and size, not from switching families. The result feels approachable without the clinical precision of a neutral sans used alone.

Colors

The palette centres on a desaturated sage green as primary — warm enough to avoid the coldness of blue-based trust signals, neutral enough not to read as a nature or wellness brand. The coral accent is used sparingly: energy without alarm. Both sit on a warm cream neutral that keeps the interface from feeling clinical. Each decision was made against a specific alternative, not just toward an aesthetic.

Illustration

Illustrations in Compawnion serve a specific function: they humanize moments of uncertainty. The illustration style is deliberately simple: flat, line-based, with limited color drawn from the product palette.

Copy

Direct, warm, without lecturing. Short sentences that meet Maja where she is, not where she should be. No pressure, no promises that are too big. Compawnion's copy follows one underlying approach: encouragement through honesty, not through optimism.

Matching System

The Alltagspassung score communicates four states — sehr gut, gut, möglich, gering. Two layers carry the meaning: colour and expression-based icons. Colour alone creates ambiguity in the middle range; the icon adds a second signal that reads independently. The result informs without requiring interpretation.

Contextual reassurance

The Tipp component appears at two specific moments: when a low match score could discourage contact, and immediately before sending a request. It is deliberately understated — visually separate, never blocking progress. The placement is a direct response to the core research finding: contact anxiety, not lack of interest, is what breaks the adoption process.

Progressive Disclosure

The accordion pattern in the match breakdown separates assessment from explanation. Maja sees the rating first and decides herself whether she wants the detail. She is not forced to process everything at once — a deliberate accommodation for a user who tends to overthink.

Testing & Iteration

The concept was understood.
The details needed work.

The transition from mid-fi to high-fi exposed the first concrete problem. Before any user was involved, a main decision from the mid-fi prototype needed revisiting: the colour logic for match quality.

What Changed

Matching colours switched from brand accent to traffic light: The mid-fi used the product's coral accent across all match levels. Participants read similar colours as similar quality. The fix: a traffic light system — green, amber, red — where familiarity carries the meaning.

Dot markers replaced with expression-based icons: Neutral dot markers were replaced with expression-based icons, adding a second layer of meaning that reads faster and requires less interpretation.

Before

After

Four remote usability tests followed with the high-fi prototype. The concept held. What surfaced were friction points in interaction and language — the kind that only become visible when real users move through a flow for the first time.

What Worked

Overall Design: Visual design, colour palette, and information density were consistently praised.

‘Alltagspassung’ Indicator: The traffic light system for daily life compatibility was understood and not perceived as discouraging.

Profile interaction: Profile features like the favourites function were used intuitively as expected.

What Changed

‘Alltagspassung’ not spontaneously understood: No test participant understood the term without context. The instinct to replace it was resisted — 'matching' carries dating-app connotations that undermine the product's tone. The hypothesis: the concept needs onboarding, not a better label. Recent feedback seems to confirm this.

‘Alltagspassung’ needs an entry point: An overlay was added to the animal list screen — the first encounter with the term after onboarding — to establish the concept before users rely on it.

Before

After

Minor Changes

Unexpected clicks: Users interacted with some elements differently than expected or had difficulty doing so. Many of these issues were resolved by adjusting or expanding the touch area.

Missing metadata: Age missing in the animal list, size in the animal profile. Both added.

Confusing CTAs: The copy caused confusion: is the message for the animal or the shelter? The ambiguity pointed to a structural problem — the app mediates between user, animal, and shelter, and the CTA sat at the intersection of all three. The solution: anchor the language explicitly to the shelter at key decision points.

AI in the Design Process

Using AI to ‘learn slower’.

I used AI tools deliberately, with the goal of understanding the design process itself, not accelerating it. Where I used AI, it was as a thinking partner, not an execution tool.

I used Claude throughout as a sparring partner: to structure research material and data, to iterate on question and answer formats in the Alltagscheck, and for the writing process behind this portfolio page. Suggesting formulations, reasoning through them, discarding them, the final decision was always mine.

I used Figma Make, Lovable and Claude Design selectively for individual screens, to quickly generate and compare variants, as a tool for decision-making, not for design.

I used ChatGPT to generate placeholder illustrations for key screens and flows. These are intentional stand-ins , the final product would use proper vector-based illustrations.

The deliberate restraint had a reason: I wanted to make decisions myself, make my own mistakes, and actually learn. It’s slower. But it gave me a much better understanding of why a design works, not just whether it does.

AI in the design process, placeholder image

Learnings & Outlook

What I’d do differently,
and what remains open.

Without shelter data, matching stays a hypothesis.

Compawnion works only as well as the data it receives. The user side is built, but without structured animal profiles from shelters, matching remains a hypothesis. This was a deliberate scope decision, not an oversight.

‘Alltagspassung’ needs onboarding, not a better label.

‘Alltagspassung’ was not spontaneously understood by any test participant. That’s not a wording problem, it shows that a new concept needs onboarding, not just a better label. Whether ‘matching’ would be clearer was discussed, but the term carries too strong a dating connotation. The hypothesis that onboarding will carry the term remains to be tested.

Research was strong early. Testing earlier would have been stronger.

The conceptual foundation was well-supported by survey and interview data. What I’d do differently: earlier testing sessions with an unfinished prototype. I waited too long for the product to feel ‘test-ready’, rough, early tests would likely have been more revealing than later ones with a polished mid-fi.