A jewellery store with hundreds of SKUs across multiple metal types and purities cannot afford a pricing workflow built on spreadsheets and CSV imports. Every time gold moves, the clock starts ticking — and the manual cycle is long, error-prone, and invisible to the customer until something goes wrong.

• ⏱ 1,200+ Shopify variants each requiring price updates 3–4× daily. The existing workflow: download CSV → update in Excel → re-upload via Shopify bulk import → verify. Each cycle took 1.5–2 hours.
• 💾 Bulk CSV imports caused random variant data corruption — weight values reset silently, discovered only after customer complaints about wrong prices.
• 🧮 No purity-adjusted pricing in Shopify. 22K and 18K gold were both managed with flat price tags — not formula-adjusted for purity. Margins on 18K pieces were systematically understated.
• 📡 The store owner was checking Mumbai gold rates from a WhatsApp group and updating prices by feel — not from authoritative MCX spot rate feeds. Accuracy was unreliable.
• 🚨 No safety net: if a formula error caused a price to drop by 50%, it would go live with no alert. The store had twice sold pieces below cost before catching the error manually.

KLYX proposed an embedded Shopify App Bridge extension (inside the merchant's existing Shopify Admin) rather than a standalone portal. Two architectural decisions drove the design: (1) a formula-based pricing engine where each product category has a pricing rule expressed as variables — metal_type, purity_code, weight_grams, making_charge_pct, wastage_pct — and (2) a variance guard that flags any variant whose new price deviates more than 8% from the previous price before the sync commits, preventing formula errors and abnormal rate spikes from going live unreviewed.

The embedded approach was deliberate. A standalone portal would require a context switch every time the owner wanted to run a sync — and that friction alone would kill adoption within weeks. Built inside Shopify Admin, the tool feels like a native feature from day one.

Once per category formula is defined, every subsequent price sync is a single click. No spreadsheets. No CSV uploads. No silent data corruption.

1. Audited the 1,200-variant catalog. Found 40% of variants had missing or inconsistent weight data. Cleaned and normalised before any automation — garbage in, garbage out.
2. Defined the pricing formula schema: each variant maps to {metal_type, purity_code, weight_grams, making_charge_pct, wastage_pct}. Making charge and wastage varied by category (bangles vs. chains vs. rings).
3. Integrated live MCX gold rate feed (10g 24K basis) with purity conversion: 22K = 24K × 0.9167, 18K = 24K × 0.75. Silver from LME basis, converted to INR.
4. Built the Formula Builder UI (Shopify Polaris components) — the owner defines per-category pricing rules without code. Changes apply to all variants in that category on next sync.
5. Replaced CSV bulk import with Shopify's GraphQL productVariantsBulkUpdate mutation. 1,200 variants now update in under 45 seconds vs. the 2-hour CSV cycle.
6. Added variance guard: if any variant's new price deviates more than 8% from its previous price, it's flagged for manual review before the sync commits. Catches formula errors AND abnormal rate spikes.
7. Preview dashboard shows every variant's current price, new calculated price, and delta (INR and %) before any changes are committed. Owner can override individual variants.
8. Automated WhatsApp summary (WhatsApp Business API) sent at 10 AM daily: current gold rate, update status, and count of flagged variants requiring review.

• Variance guard caught 3 formula errors in the first month before they went live.
• Owner stopped managing a WhatsApp group for gold rates — MCX feed replaced it entirely.
• Preview dashboard gave the team confidence to run Auto-Approve within 2 weeks of launch.

• Data quality before automation: before automating any pricing system, normalise your variant data. 40% of this store's variants had corrupted weight data — any formula applied to bad data produces bad prices. Run a catalog audit first.
• The variance guard is underrated: a simple rule that flags >8% price changes catches formula errors AND abnormal market movements before they reach customers. Every pricing automation should have this safety layer built in from day one.
• WhatsApp summaries close the loop: the daily summary gave the owner confidence that automation was running correctly without requiring a Shopify login every morning. That trust is what makes automation stick long-term.