Dekel Factor Designs
Case study · AI fashion e-commerce

Season

Confidence looks good on you.

Color intelligence, virtual try-on and wardrobe building, designed around the person wearing the clothes.

Role
Product Designer
Timeline
8 months
Platform
Responsive web · desktop and mobile
Scope
Color match profile, virtual try-on, outfit builder, AI stylist

3D is unavailable in this browser, so the case study shows without its animated backgrounds.

Skip to final design
Scroll the story
01 / The brief

Does itworkon me?

Color mismatch, uncertain fit and endless choice make online shopping feel like guesswork. Season turns those anxieties into product decisions.

Season explores how AI could read a person’s color signature from a photo, match outfits and retailers to it, and end in a fitting room where they see the look on themselves before they buy.

My role

What I owned,and with whom.

Role
UX/UI & product designer
Client
Season
Timeline
8 months
What I did
ResearchUX flowsUI designPrototypeBuild
02 / Research & goals

Listen beforechoosinga palette.

The study combined reports with interviews with online shoppers. The goal that came out of it: guidance without dictation. Every conversation pointed to one of three needs.

“Colors look different on my skin tone than on the site.”
“I always end up returning half of what I buy online.”
“I wish I could see how an outfit looks on me before buying.”

What shoppers said · from the empathy map

01 / COLORUncertainty.

A screen can’t explain how a shade relates to the shopper’s undertone.

02 / FITDoubt.

People want a sense of drape and proportion on their own body.

03 / CONTROLTrust.

Personalization should feel explainable and adjustable.

02 / Existing alternatives

Three services.Pieces ofthe answer.

The competitive review drew on Nordstrom, Wishi and Style DNA. The distinction Season aims for: combine color intelligence with user control and access to human styling.

NordstromWishiStyle DNA
Color matching360° try-onStyling guidanceWardrobe planningTrust + privacy
DESIGN OPPORTUNITY

From palette to purchase.

Turn a color analysis into outfits people can try on and buy, in one continuous flow.

02 / Feature comparison

Everyone had a piece.No one had the path.

Five capabilities across Nordstrom, Wishi and Style DNA. Only Style DNA matches color; none offers a realistic try-on or wardrobe planning.

Feature comparison of Season, Nordstrom, Wishi and Style DNA
CapabilitySeasonNordstromWishiStyle DNA
Color match (undertone-aware)Season✓Nordstrom✕Wishi✕Style DNA✓
360° realistic try-onSeason✓Nordstrom✕Wishi✕Style DNA✕
Styling guidance / outfitsSeason✓Nordstrom◐Wishi✓Style DNA✕
Saved profile + wardrobe planningSeason✓Nordstrom✕Wishi✕Style DNA✕
Trust + privacy controlsSeason✓Nordstrom✕Wishi◐Style DNA✕
✓ Has it◐ Partly✕ Missing
03 / Who it’s for

Meet Noa.

Portrait of Noa Azulai

Noa Azulai

26 · Fashion enthusiast · Tel Aviv

“Colors look different on my skin tone than on site. I end up buying “safe” items only because styling takes too long.”

Confident shopping, zero guesswork. A busy professional who shops mostly online, loves fashion, but hates wasting time and returning items. She needs a reliable way to find pieces that flatter her colors and shape.

GOALS
  • Quickly find pieces that flatter her colors and shape
  • Build cohesive outfits from fewer items
  • Reduce returns and decision fatigue
  • Feel confident buying online without try-before-buy
PAIN POINTS
  • Colors look different on her skin tone than on site
  • Not enough confidence to buy higher-priced items online
  • Styling takes too long, so she buys “safe” items only
  • Orders multiple sizes and colors “just in case”

The main persona, drawn from user interviews

03 / User story

Same outfit.Two different carts.

Noa’s goal never changes: find pieces that suit her and buy them with confidence. What changes is how much guesswork the store leaves her to do on her own.

Without the app

Fragmented path

  1. 01
    Guess the colors

    Product photos look different on her skin tone.

  2. 02
    Order three sizes

    Buying “just in case” because fit is uncertain.

  3. 03
    Style it alone

    Putting looks together takes too long, so she buys safe.

  4. 04
    Send most of it back

    Returns cost time, and the confidence doesn’t grow.

Buying safe, returning the rest

With Season

Connected path

  1. 01
    Match her colors

    Shades are checked against her own skin tone.

  2. 02
    Try it on in 360°

    Fit and drape are visible before she orders.

  3. 03
    Get styling guidance

    Looks are suggested around pieces she already owns.

  4. 04
    Buy once, with confidence

    Her profile and wardrobe are saved for next time.

Buying once, with confidence

One color profile

Shades that suit her, every time

One try-on

Fit is seen, not guessed

One wardrobe

New pieces work with what she owns

04 / User flow

From a first phototo a look she buys.

Before the screens had a look, they had a structure. Begin the journey opens three doors. Uploading photos runs through sizes and analysis; signing in or choosing a model skips both, since the profile or the model already holds that data. All three meet at the palette, and in the fitting room an overlay swaps the model.

ScreenChoiceShortcut

04 / User journey

A lookof your own.

Each stop on the line is a screen. Scroll to follow Noa from a first photo to a look she’s ready to buy.

01 / 05Curious

Starting point

She chooses how to begin: her own photos, or a model close to her.

DesignTwo doors into the same flow, so privacy never blocks the experience.

02 / 05Hesitant

Photos & model

She adds two daylight photos, guided by a simple frame.

DesignClear photo guidance, and a model option for anyone who’d rather not upload.

03 / 05Focused

Fit & style

A few measurements set her fit.

DesignOnly what fit needs; everything else is optional.

04 / 05Delighted

Color signature

The orb blooms into her palette: her season and her best tones.

DesignThe reveal is the emotional peak, so it gets the whole screen.

05 / 05Confident

Fitting room

She tries Look 01, switches color, size and angle, and sees where to buy it.

DesignEvery control on one screen, ending at a real retailer.

Into your colorsNow, the fitting room.

05 / Final design

Closer to the product.Closer to the decisions.

Scroll through the flow. Each state carries one design decision.

  1. 01 Start

    Let people choose the door.

    Photos or a model: both paths lead to the same fitting room.

  2. 02 Photos

    Guide the photo, not the person.

    A frame and three light tips replace a wall of instructions.

  3. 03 Fit & style

    Ask only what fit needs.

    A short sizing step keeps momentum toward the reveal.

  4. 04 Your edit

    Make the reveal the moment.

    The color orb, faint through onboarding, blooms here into her palette.

  5. 05 Fitting room

    Try, adjust, buy.

    Outfit, size, color, model and angle on one screen, with retailers to buy from.

The flow, screen by screen

Five screens. One flow.

The final design, desktop and mobile, side by side. Scroll or drag sideways to turn through the flow.Scroll or swipe sideways to turn through the flow.

Season desktop. Choose your starting point: log in, use my photos, or choose a model
Season mobile. Choose your starting point: log in, use my photos, or choose a model
01Starting point
Season desktop. Build your portrait set: guided front, left, right and three-quarter photos
Season mobile. Build your portrait set: guided front, left, right and three-quarter photos
02Photos
Season desktop. Your fit, your style: clothing size, shoe size and fit preference
Season mobile. Your fit, your style: clothing size, shoe size and fit preference
03Fit & style
Season desktop. Your Soft Autumn edit: color signature, palette and matched looks
Season mobile. Your Soft Autumn edit: color signature, palette and matched looks
04Your edit
Season desktop. Virtual fitting room: model in the look, colorway, sizes and retailer choice
Season mobile. Virtual fitting room: model in the look, colorway, sizes and retailer choice
05Fitting room
The full flow, in motion

From your palette to your pieces.

The live Season flow end to end: a portrait set and a fit profile become a personal palette, every colour in it pulls the looks that match, and the chosen look goes to the fitting room, the bag and the retailer checkout.

06 / Decisions & trade-offs

Every screenwas a fork.

The three calls that shaped the product most, what I chose, and what I let go.

Fork 01 · Photos

Require a photo, or not?

Chosen

Offer both: your own photos, or a model matched to your proportions.

Rejected

Requiring a personal photo before anything else.

Why. A mandatory selfie is the biggest drop-off point for a styling product. A model path keeps privacy-conscious people in the flow.

Evidence · 63% expect personalization to feel accurate, transparent and controllable.

Fork 02 · Onboarding

Overlay or full screen?

Chosen

Full-screen steps, one decision per screen.

Rejected

The original overlay modal with Models, Photos and Measurements tabs.

Why. The overlay hid the product behind a form and packed three tasks into tabs. Full-screen steps give each decision room and make progress visible.

Evidence · Competitor review: Wishi mixes mobile and desktop UI into confusing navigation, and Nordstrom reads as cluttered.

Fork 03 · Paywall

How free is free?

Chosen

A soft paywall: free try-ons first, then an upgrade to premium AI engines.

Rejected

Locking the fitting room behind an account from the start.

Why. People need to feel the value before paying, so the lock appears only after they’ve seen themselves in a look.

Evidence · 72% say try-on experiences increase purchase confidence.

07 / The store around it

The fitting roomopens intothe store.

The home experience speaks the same color language: tonal discovery, categories, top picks and retailers, all filtered through a person’s palette.

Tonal discovery

Products grouped by tone, not just by category.

Shop by category

Categories re-ranked to her palette.

Who wore it best

Social inspiration in the same color story.

Top retailers

One place to find a look across stores.

From orb to system

One orb.Four tokens.

The color orb that blooms through the flow is where the palette comes from. Primary, signal, resolve and surface leave it one by one and settle into the system.

08 / Style guide

Room to see.Room to choose.

Instrument Serif and DM Sans pair with a near-black night, warm ember and earthy seasonal hues. Desktop and mobile use layouts suited to their own interactions, and a recurring color orb grows from faint to full bloom across the flow.

Identity
Season logo: a glowing S-shaped orb in warm ember tones beside the Season wordmarkSeason

A glowing S-shaped orb in warm ember tones beside a wordmark set in Instrument Serif.

Typography
DisplayInstrument SerifAaDressed in your colours.Aa Bb Cc Dd Ee Ff Gg Hh 0123456789
BodyDM SansAaDiscover pieces that match your colour signature.Aa Bb Cc Dd Ee Ff Gg Hh 0123456789

Instrument Serif for editorial names and headlines. DM Sans for reading text and controls, at 600 and 400.

Colour
Ember#DF8068

Primary

Wine#732536

Signal

Sage#5D6D61

Resolve

Cream#F2EADF

Surface

Components

Rebuilt from the live Season site: Archivo and DM Sans controls, Instrument Serif names, ember accents and squared edges. Everything below is live, so try it.

Buttons
Text links
Start your analysis ↗View All Collections ↗
Search & browse
Product card
Olive bomber jacket with white tee, grey jeans and black high-tops
Bomber Jacket$189
Category & community vote
Black leather jacketOuterwear42 pieces
Street-style outfit with sunglassesLeading
Retailer selector
01One Store

Buy the entire look from one retailer, in one checkout.

4 pieces · 1 checkout$496
Palette chips
MerlotTerracottaOchreSageMossIvory
Sign-up form
09 / Accessibility & trust

Color,without exclusion.

A product built on color has to work for people who see color differently, and for people who don’t want to share a photo.

Never color alone

Every swatch has a name and every palette a written description, so color-blind users get the full result.

Photo privacy

Photos are used for analysis only, with a clear option to delete them, and a model path that needs no photo.

Text over imagery

Labels on product photos keep WCAG AAWeb Content Accessibility Guidelines, level AA — the standard for colour contrast and readability. contrast with a scrim, not luck.

Calm motion

The breathing orb respects reduced-motion settings and never flashes.

10 / Projected outcomes

What the researchpointed toward.

These are targets estimated from the research, not tested results. The proposed measurement points are profile completion, try-on success, saved outfits and repeat visits.

Estimates · research-informed targets, not tested results

≥70%

reach their color signature after starting onboarding

Target · completion
−30%

color-related returns for shoppers who use the fitting room

Target · returns
<2 min

from first photo to palette reveal

Target · time on task
3 / 3

research needs covered: Read, Match, Try

Design coverage

How these were estimated

Each target traces back to a finding from the research: colors that disappoint in person, overwhelming catalogs, and no way to try before buying. The research behind them: 56% of fashion returns stem from color or fit mismatch, 72% say try-on increases purchase confidence, and 63% expect personalization to be accurate, transparent and controllable. A moderated test on the prototype would be the first step to confirm them.

11 / Reflection & next steps

What worked.What I’d change.

What worked

A recurring motif as a guide.

Carrying one color orb through the whole flow, faint at first and blooming at the reveal, gave the journey a visible arc.

What I’d change

Validate the color analysis itself.

The whole experience rests on the palette being right. I’d test the AI’s results against people’s own judgment, and against stylists, early.

What I learned

Design the prompt, not only the pixels.

I locked every visual decision in ChatGPT first, so the Figma Make build became one precise execution pass instead of rounds of trial and error.

Where I would take it next

Next: deeper hair, eye and skin analysis, A/B testing trust signals, the soft-paywall and account flow, and the new design language on every page.

This is a design case study; the outcomes above are research-based estimates, not measured results.

Season · live site