Archetipas · Shopify e-commerce
Visit live site ↗
archetipas.lt
ROLE
Designer & Shopify developer
TIMELINE
Summer 2026 · live 22 Aug 2026
STACK
Figma, Shopify OS 2.0, Liquid, CSS, JS
SCOPE
Audit, design, build, handoff, review
MY ROLE
I ran the audit, designed everything in Figma, built it as a custom Shopify theme, handed it over with a guide, and came back after launch to measure what happened.
I OWNED
•
Strategy audit and goals
•
Full design in Figma
•
Custom Shopify OS 2.0 theme (Liquid, CSS, JS)
•
Metaobjects, SEO and structured data
•
Client guide and handoff
•
Post-launch analytics review
THE CLIENT OWNED
•
Content, products and photography
•
Ad creative and budget
•
Domain and payments
TOOLS
Figma
Shopify OS 2.0
Liquid
CSS / JS
Metaobjects
JSON-LD
Hotjar
from ~80%
bounce rate after launch
from ~2%
conversion, under scaled paid traffic
from 14–25 a month
orders a month after launch
Sources: Shopify, Hotjar and Contentsquare windows before and after 22 Aug 2026. Revenue is not shown.
THE CHALLENGE
•
The problem was not “more ads”. It was trust and a clear path to purchase.
•
Deep, reflective products need a calm, credible place to be sold.
•
The goal: about 600 orders in 6 months, roughly 100 a month.
CONVERSION
ORDERS / MONTH
BOUNCE
PAGES / SESSION
STRATEGY
Traffic is not conversion
Ad spend without a store that sells just burns budget. Fix the store first.
Mobile first
Almost every shopper arrives on a phone, from an ad. Design that path: ad, product, checkout.
Custom theme, not a tweak
The brand needed its own visual language, so I built from Figma instead of restyling a stock theme.
The client stays independent
Content, variants and SEO fields are editable without me. No designer needed for a text change.
HOW I GOT THERE
01 · RESEARCH
Before drawing anything I wrote a strategic audit from Shopify and Meta data. The funnel showed plenty of views, 4.24% added to cart and 2.04% bought. The buyer is analytical and sceptical, so the concept became “structured chaos”: a strict grid with real, raw content.
DECISION
Fix trust and the path to purchase. Leave the ad budget alone.
Pages from the audit I presented to the client (in Lithuanian).
Four type directions for the same hero. The outlined one won.
02 · TYPOGRAPHY
I set the same headline in four directions, from loud script to elegant display faces, and checked each against real-world references.
DECISION
A classic serif. Calm and credible beats loud for this buyer.
03 · LOGO
Dozens of wordmark tests: capitals, stacked letters, circles, a full red sun. Most of them said more than the brand needed.
DECISION
A lowercase wordmark with one red dot on the i. The dot ties back to the cover.
Six directions from the logo exploration, and the one that shipped.
Six background and accent combinations, with the palettes they came from.
04 · COLOUR
I tested the hero on sand, teal and off-white with red, blue and green accents, next to palettes pulled from the moodboard. Every extra colour took attention away from the product.
DECISION
One accent, burgundy. Everything else stays neutral.
05 · FIRST DRAFT
The first homepage used a moody product crop and uppercase everywhere. Once the client’s own photography came in, the page felt warmer with people in it and lighter type.
DECISION
Real photography, one promise, one red button.
An early homepage draft. Compare it with the shipped hero below.
PROCESS
01
Audit & goals
Where visitors drop and what success means.
02
Figma system & pages
Tokens, components and every template.
03
Custom theme
Shopify OS 2.0, sections and Liquid.
04
Product UX & SEO
Variants, metaobjects, structured data.
05
Publish
Live on 22 Aug 2026, baseline metrics locked.
06
Client guide
“Kaip naudotis”: how to run the store.
07
Live fixes
Mobile gallery overflow and other details.
08
October review
What the data says, and a 3-step plan.
DESIGN LANGUAGE
TYPE
Aa Aa
Cardo for display, a clean sans for body. Colours and logo are exposed in theme settings, so the brand can evolve without code.
COLOUR
Burgundy
#AF2928
Near-black
#141414
Charcoal
#3A3A3A
White
#FFFFFF
WHAT I DID
01
One strong image, one promise, one red button. Visitors from an ad understand in seconds what this is and where to go.
02
Gallery, price and edition cards side by side. The standard and the discounted print-error edition are shown as clear, comparable choices.
WHAT WAS BUILT
12 templates
Home, About, Initiative, Product, Quiz, Blog, FAQ, Contact, Cart, Collection, 404 and accounts.
Edition cards
Variant picker as cards: standard cover versus print-error edition.
Self-sabotage quiz
20 questions as a soft way into the product.
Metaobjects
Partner logos and reviews managed in one place, shown on many pages.
SEO foundation
Meta fields and JSON-LD for products, articles, FAQ and breadcrumbs.
Sticky header
Hides on scroll down, returns on scroll up.
ON MOBILE
Home, quiz and product page on mobile. After launch I fixed the product gallery so thumbnails scroll inside it and the page no longer shifts sideways.
RESULTS
Bounce rate
Before
~80%
After
~63–67%
Pages per session
Before
~1.45
After
~2.2
Session time
Before
~1:00
After
~1:20+
Conversion
Before
~2–2.5%
After
~3.2%
Orders per month
Before
14–25
After
90–100
Largest contentful paint is about 1.0 s on recent measurements.
ORDERS
1–21 Aug
Old theme. Ads already running.
22 Aug – 19 Sep
New theme live, ads widened to more regions.
From 20 Sep
The winning ad scaled up.
The windows differ in length, and ad spend grew in the same period. The store’s part is the conversion rate: it rose from about 2% to about 3.2% while traffic scaled.
HANDOFF
CLIENT
•
Content, products, variants
•
Menus and SEO fields
•
Ad creative and budget
•
Domain and payments
ME
•
Theme bugs
•
Agreed post-launch review
•
Small data-informed fixes
EXTRA, QUOTED
•
English and markets setup
•
Email automations
•
Branding phase II
THE OUTCOME
•
Separate what ads fix from what the store fixes, and measure both.
•
Compare absolute rates, not flattering year-over-year percentages.
•
A written handoff with clear boundaries keeps a small project healthy.
Thor’s take
Conversion went from 2% to 3.2%. My conversion rate on begging at the table is 100%. I could consult.
NEXT PROJECT
