A D2C jewelry storefront with story behind every piece.
Gold-plated jewelry across four narrative collections — Evil Eye, Serenity, Bloom, Bohemia. Each got its own editorial voice. None broke checkout.
- Client
- Soulista
- Sector
- Jewelry · D2C E-commerce
- Engagement
- Theme rebuild → Launch
- Year
- 2026
✱ § 01 · The brief
What they actually needed.
Soulista sells gold-plated jewelry across four named collections, each with its own meaning — protection (Evil Eye), peace (Serenity), growth (Bloom), self-expression (Bohemia). The brand voice on Instagram was editorial and intentional. The Shopify storefront was a generic theme that treated every collection identically.
73% of sessions arrived from Instagram. Those visitors expected the editorial energy of the feed; they landed on a templated grid with the same product page, the same cross-sell, the same hero. The brand's voice didn't survive past the homepage — and AOV reflected it.
“An Evil Eye buyer is not a Bohemia buyer. The site treated them like the same person.”
✱ § 02 · Our approach
Three bets, in order.
Collection pages as editorial moments.
Each of the four collections gets a custom hero, a story spread (origin, symbol, intention), and a curated grid. Not a templated feed with swapped photography — a different layout per collection, sharing only the product card.
Bundle dynamic on every product page.
Three 'complete the look' suggestions weighted by collection affinity, not just collaborative filtering. An Evil Eye necklace recommends Evil Eye earrings before it recommends a generic best-seller.
Klaviyo flows tuned per collection.
Replaced the single welcome flow with four. An Evil Eye first-buyer doesn't get a Serenity drop email; the messaging matches the intention they bought into. Segmentation by collection from purchase one.
✱ § 03 · Process · 9 weeks
From kickoff to shipped.
Wk 1
Brand audit + Instagram analysis
Mapped Instagram engagement by collection vs. site conversion by collection. Found the gap — Evil Eye had 4× the IG saves, half the on-site CR. Brand voice loss confirmed.
Wk 2–3
Custom theme architecture
Designed the collection-page template family — shared header / shared product card / different middle sections. Photography brief for the four story spreads.
Wk 4–7
Build · theme + Klaviyo
Custom Shopify theme with four collection page variants, dynamic bundles via Search & Discovery, four Klaviyo flows. Performance budget: < 1.8s LCP on mobile.
Wk 8
Soft launch + photography
Cut over Evil Eye and Serenity first. Watched IG → site → checkout funnel for two weeks. Fine-tuned bundle weights — Evil Eye over-recommended itself, dropped to 2/3 same-collection, 1/3 cross.
Wk 9
Bloom + Bohemia + handover
All four collections live. Trained the in-house team on collection-page CMS edits and the per-collection Klaviyo flow. Launched a gift-wrap option as a phase-two add.
✱ § 04 · Gallery · key screens
Selected screens.
✱ § 05 · Results · measured 6 months post-launch
What changed, in numbers.
- 1.8s
- Mobile LCP
- +94%
- Collection page CR
- AED 312
- Average order value
- +56%
- Instagram → checkout
Down from 4.9s on the legacy theme
Average across the four collections
Up from AED 178 baseline
Completed-checkout attribution
✱ § 07 · In their words
The site finally feels like our Instagram. Customers spend more because the story doesn't break at the product page.
Founder
Soulista · Soulista
✱ § 08 · Tech stack
What's running.
- Storefront
- Shopify · Custom Liquid theme
- Recommendations
- Shopify Search & Discovery · custom weights
- Lifecycle
- Klaviyo (4 collection flows)
- Payments
- Apple Pay · Card · Tabby
- Analytics
- GA4 · Shopify Analytics · Hotjar
- Hosting
- Shopify · Cloudflare
✱ § 09 · Team
Who built it.
- Affan Manzoor
- Project lead
- ChampX design
- Theme + collection art direction
- ChampX engineering
- Liquid + bundles
- ChampX content
- Collection story copy
✱ § 10 · Honesty section
What we’d do differently.
01 · Custom theme made future-proofing harder
Bespoke Liquid means non-trivial Shopify version updates. Wrote a clear handover doc and a quarterly upgrade plan, but next time we'd lean on Hydrogen or Online Store 2.0 sections more.
02 · Bundle weights needed manual override
New-drop products had no purchase history, so the algorithm under-recommended them. Added a manual 'pin to bundle' toggle in the admin for the first 30 days of any new product.
03 · Should have shipped gift-wrap with v1
Late November traffic spiked and we hadn't shipped gift-wrap as a checkout option. Added it in week 14 — it became 18% of checkout adds within a month.
