Miro is a personal styling software that leverages principles of adaptive UX and recognition of physical features to support users in choosing personalised clothing and make-up.
Its goal is to improve self-perception through a digital experience that is inclusive, accessible and data-driven, combining aesthetics, simplicity and technology.

Problem

People dedicate less and less time to shopping and often buy garments that do not flatter their body shape or end up unused.
Difficulty in recognising suitable colours and shapes leads to waste, impulsive choices and low purchase satisfaction.

Solution

Miro helps users to optimise their wardrobe and purchases through a system of personalised advice based on objective data: analysis of body proportions and colour harmony.
The experience is designed to reduce the average decision time and increase the accuracy of recommendations based on individual characteristics.

The process

Discover

Interviews, desk research, competitor analysis → definition of problem/objectives.

Define

Personas, empathy map, information architecture and user flow.

Ideate / Prototype (lo-fi)

Wireframes and low-fidelity prototype; quick tests on critical flows.

Test

Visual & Prototype (hi-fi)

Visual design, high-fidelity prototype; moderated usability testing.

Test

Iterate & Handoff

Iterations based on insights, design system and handoff to development.

Research

Goal

Understand attitudes, needs and friction points in clothing/make-up choices, in order to define personalisation criteria and product priorities.

Mixed methodology

Quantitative questionnaire + qualitative interviews.

Quantitative research

Sample

50 potential users (screened: age 25–45, purchasers in the last 6 months).

Tool

Online questionnaire (Google Forms).

Note

Indicative results (not representative of the entire population); useful to guide design hypotheses.

“Do you ever feel you have nothing to wear even though you own many clothes?”


“sometimes” 55%

“often” 40%

“never” 5%

Shows high potential for everyday decision support.

“Are you interested in improving your style?”


“quite” 75%

“a lot” 17%

“not interested” 8%

Strong openness to personalised advice.

“Do you own clothes you have never worn?”


“some” 57%

“many” 29%

“few” 8%

“none” 6%

Opportunity: reduce waste and unfocused purchases.

Quantitative insights
High motivation to improve

Room for recommendations and light coaching.

Perceived waste

Value proposition around more mindful purchases.

Frequent indecision

Need for quick guidance.

Qualitative research

Sample

10 semi-structured interviews (5 women, 5 men, 28–45).

Focus

Choice process, aesthetic/functional criteria, in-store/online friction points, self-perception.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

  • They want targeted purchases, less impulsive; they value eco-sustainability and quality.

  • They prefer trying on in store to avoid wrong sizes/fit, but they look for pre-selections online.

  • Curious about trends, but with a preference for tailored advice and personal growth.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

  • Personalised suggestions for colour/morphology with user control (non-prescriptive).
  • Reduced decision time with smart shortlists and “why we recommend this” explanations.
  • Digital pre-fitting (sizes/fit) and wardrobe management to limit unused purchases.
User Personas

Research data and insights were synthesised into two key profiles that represent the main behaviours and needs of target users.

The personas guided the definition of user flows, content and the visual tone of the app, helping to balance efficiency and inspiration in the digital experiences.

Chiara – Efficiency and precision

Beatrice – Creativity and spontaneity

The two main personas, Chiara and Beatrice, represent the profiles that emerged from the research.
They guided the definition of the app flows and visual tone: one faster and more functional, the other more exploratory and inspirational.

Starting from their goals and digital habits, I defined the main usage flows, which were then tested with low-fidelity prototypes.
MVP

Version tested to validate the interest in and effectiveness of the concept

Core features
  • Collection of images of the garments owned by the user.
  • Creation of outfit combinations starting from the uploaded garments.
  • Automatic analysis of colour harmony, body shape and facial features.
  • Generation of a list of recommended garments based on the analysis results.

The MVP focuses on the most immediate experience for the user: seeing their wardrobe digitalised and receiving personalised suggestions.
This first version was designed to test the technical feasibility of the visual analysis and the perceived usefulness of the recommendation system.

Flow

Definition of the main paths for the MVP

Main flows mapped:
  • Onboarding and login
  • Profile creation (colour analysis, body shape, face)
  • Item upload and outfit creation
  • Analysis and recommendations
Flow goals
Make the relationship between features and user goals clear
Reduce unnecessary steps (e.g. from 7 to 4 steps to create an outfit)
Validate the logic of the journey before the prototype
Wireframes

First low-fidelity layouts to validate flows, hierarchies and micro-interactions before the visual design.
The pink areas indicate the tested hotspots; the numbering follows the main journey (1→10: Home → Analysis → Shopping → Outfit → Wardrobe → Favourites → Details).

Key results
  • Reduced the steps needed to create an outfit (from 7 to 4)
  • Clarified the primary CTA
  • Made upload and “save outfit” more visible.
UI design

High-fidelity interface designed to balance operational speed (cards, 5-item bottom nav) and visual inspiration (grids, outfit previews).
The hierarchy is driven by clear CTAs, consistent components and micro-feedback.

Design choices
  • Card + grid: fast browsing, focus on outfits.
  • Bottom navigation: 5 fixed sections (Home, Features, Shopping, Outfits, Wardrobe) to reduce click depth.
  • Primary CTAs (“Save your outfit”, “Discover”) with strong contrast and clear state; lighter secondary CTAs.
  • Tags/filters at the top right for immediate control over recommendations.
  • Soft illustrations and a neutral palette for an inclusive, non-prescriptive tone.
Adaptive UI

Colour analysis and personalisation

Concept

The interface adapts to the user’s colour profile (colour analysis) identified in the initial test.
Depending on the user’s season — Winter, Autumn, Summer or Spring — palettes, background tones and illustrations change while remaining consistent with the brand system.

User experience

After the test, the user sees their Home personalised with colours that are harmonious with their complexion and visual traits.
The goal is to increase empathy and sense of recognition, turning the UI into an extension of the person.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

I am text block. Click edit button to change this text. Lorem ipsum dolor sit amet, consectetur adipiscing elit. Ut elit tellus, luctus nec ullamcorper mattis, pulvinar dapibus leo.

“