Shopify Plus Migration and Paid Media Optimization in Pakistan

The Client

A Karachi-based D2C clothing brand launched in 2020, targeting women 18–34 across Pakistan’s urban centers with affordable unstitched fabrics, pret wear, and seasonal collections. The brand had grown to PKR 8.5M/month in revenue (70% from paid channels, 30% from organic/social) and was processing 1,200–1,500 orders/month through a hybrid of cash-on-delivery (55%) and prepayment (45%). The tech stack was Shopify Basic with a heavily customized theme, 20+ apps for functionalities ranging from reviews to currency conversion to SMS notifications, and a fragmented data setup with Google Analytics 4 (partial implementation) and a basic Google Ads conversion tracking pixel. The marketing team of two (one performance lead, one creative lead) was managing campaigns in-house but hitting the limits of what the current setup could support at scale. The brand had raised a pre-Series A round in 2024 and was positioning for a Series A, with investor pressure to show efficient unit economics at PKR 1.5M+ monthly revenue.

The Problem

The brand’s paid performance had plateaued. Meta and Google Ads spend had grown from PKR 800K/month to PKR 1.2M/month over six months, but ROAS remained stuck at 1.6x — below the 2.5–3.5x fashion benchmark and insufficient to justify further spend scaling. Three systemic issues emerged from the diagnostic:

• Platform debt: Shopify Basic was hitting order volume limits during peak periods (Eid, wedding season), causing abandoned checkouts and failed transactions. The 20+ app stack was creating script bloat, slowing page speeds, and conflicting during checkout. The brand had outgrown the platform but was hesitating on Plus due to perceived cost and migration complexity.

• Feed quality collapse: The product feed to Google Merchant Center was a flat file with no variant structure. Color, size, and material attributes were missing, causing 70% of products to be disapproved or limited in Shopping campaigns. Google Shopping accounted for 35% of spend but only 22% of revenue — a clear sign of wasted impressions on mismatched queries.

• Campaign architecture rot: Meta and Google campaigns were organized by ad set/creative rather than by product category or audience intent. This prevented proper budget allocation to high-margin categories (unstitched fabrics had 3x the margin of pret wear) and made it impossible to measure ROAS by product line. The brand was over-spending on low-converting audiences (cold lookalikes) and under-investing in retargeting warm pools.

Phase 1 — Platform Migration and Data Cleanup (Weeks 1–3)

The first three weeks focused on the Shopify Plus migration and resolving the technical debt that was constraining performance. This wasn’t just a platform upgrade — it was a consolidation and cleanup of the entire marketing stack. We applied our shopify marketing agency framework, which treats platform migration as a performance optimization lever, not just a technical project.

Migration actions:

• Shopify Plus implementation: Migrated from Shopify Basic to Shopify Plus with a phased cutover (test store → staging → production). Configured Shopify Flow for automation (order tagging, high-value customer routing, inventory alerts) and consolidated 8 apps into native Plus features (abandoned checkout emails, currency conversion, basic fraud detection).
• App consolidation: Audited the 20+ app stack and identified 12 apps for removal (redundant functionality, poor performance, or replaced by Plus features). Removed 4 review apps (consolidated to one), 3 currency apps (replaced by Plus multi-currency), and 2 SMS apps (consolidated to a single provider). Reduced script load by 65%.
• Theme optimization: The custom theme was refactored to remove unused JavaScript hooks, defer non-critical CSS, and implement lazy loading for product images. Image compression was applied to the entire catalog (reduced average image size from 380KB to 95KB without visual quality loss).
• Checkout enhancement: Implemented Shopify Script Editor for custom checkout logic (free shipping threshold, gift with purchase, Karachi vs. rest-of-Pakistan shipping rules). Added one-page checkout with address autocomplete and payment method consolidation (JazzCash, EasyPaisa, HBL,UBL all surfaced prominently).
• Tracking audit: Implemented GA4 with enhanced ecommerce (view_item, add_to_cart, purchase events) and server-side Google Tag Manager for conversion tracking. Added Facebook CAPI for server-side event deduplication and privacy-safe attribution.

Pre-Migration State	Post-Migration State
Shopify Basic (order volume limits)	Shopify Plus (unlimited orders + Flow)
20+ apps installed	8 apps (native Plus features replaced 12)
Custom theme with 380KB average image	Refactored theme, 95KB average image
LCP 4.8s, CLS 0.35	LCP 1.9s, CLS 0.08
Basic GA4 + client-side GTM	GA4 enhanced ecommerce + server-side GTM + FB CAPI

Week 3 outcome: Migration complete with zero downtime (cutover during low-traffic window, 2:00–4:00 AM Pakistan time). LCP improved from 4.8s to 1.9s. Checkout abandonment rate dropped from 72% to 61%. The brand could now process peak-order volumes (2,500+ orders/day during Eid spikes) without hitting platform limits.

Phase 2 — Feed Optimization and Structured Data (Weeks 3–5)

With the platform foundation solid, we focused on the feed quality issue that was suppressing Google Shopping performance. This phase structured the product data for both Google Merchant Center and schema.org rich results.

Feed optimization:

• Product variant restructuring: The flat product feed was restructured to a variant-level feed where each size/color combination was a separate entry with its own id, title, description, and image_link. This allowed Google Shopping to match specific variant queries (e.g., “blue lawn size 40”) to exact products instead of generic product pages.
• Attribute enrichment: Added mandatory attributes across the catalog — color (e.g., “navy blue”, “maroon”), size (e.g., “40”, “42”), material (e.g., “cotton lawn”, “chiffon”), pattern (e.g., “solid”, “printed”), and brand (consistent brand name across all SKUs). Mapped Shopify product tags to feed attributes for automated updates.
• Google Merchant Center diagnostics: Ran feed diagnostics and identified disapproval reasons — missing gtin, mpn, and inconsistent image_link formats. Worked with the catalog team to add product identifiers and standardize image URLs (no query parameters, consistent filenames).
• Structured data implementation: Added Product schema markup to all product pages, including name, description, image, brand, offers (price, availability, seller), and aggregateRating (where reviews existed). Added Organization schema on the homepage for brand entity building.
• Feed rules setup: Configured Google Merchant Center feed rules to auto-categorize products using the existing product_type attribute and to suppress out-of-stock variants from active campaigns.

Feed Quality Metric (Pre)	Feed Quality Metric (Post)
Feed approval rate: 68%	Feed approval rate: 94%
Variant attributes present: 30%	Variant attributes present: 100%
Google Rich Results eligibility: 12%	Google Rich Results eligibility: 78%
Product-level impressions in Shopping: 45K/day	Variant-level impressions: 82K/day

Week 5 outcome: Feed approval rate rose from 68% to 94%. Google Shopping impressions increased from 45K/day to 82K/day with the same ad spend, indicating better query matching and less wasted spend. Rich Results eligibility jumped from 12% to 78%, positioning the brand for organic visibility gains. This feed optimization aligned with our shopify seo services playbook for organic traffic capture.

Phase 3 — Campaign Rearchitecture (Weeks 4–7)

With platform stability and feed quality solved, we rebuilt the paid campaigns from the ground up. The old campaign structure (by ad set/creative) was replaced with a product-first architecture aligned to margin, category, and intent.

Google Ads rebuild:

• Campaign restructuring: Created separate Shopping campaigns by product category — (1) Unstitched Fabrics (high margin), (2) Pret Wear (mid margin), (3) Sale/Clearance (low margin). Each campaign had its own budget bid strategy, with aggressive bidding on unstitched fabrics (target ROAS 3.5x) and conservative bidding on sale items (target ROAS 2.0x). We leveraged google shopping ads management best practices for margin-aware budget allocation.
• Priority bidding tiers: Implemented a 3-tier priority structure within Google Shopping — High Priority (new arrivals, bestsellers, high-margin categories), Medium Priority (seasonal collections, mid-margin), Low Priority (sale items, overstock). Budget was allocated 50%/35%/15% respectively.
• Audience segmentation: Added in-market audiences (e.g., “Wedding attire”, “Eid clothing”) as observation-only layers to identify high-intent users. Created a remarketing campaign for site visitors (past 30 days) with custom creatives showing “Previously Viewed” products.
• Negative keyword management: Built a negative keyword list to exclude non-converting queries — “free”, “cheap”, “wholesale”, “bulk order”, “international shipping” (the brand only shipped within Pakistan).

Meta Ads rebuild:

• Catalog-based campaigns: Replaced ad set-based campaigns with Meta’s catalog sales campaign structure using the Facebook Product Catalog. This enabled dynamic product ads (DPA) that showed the exact products users viewed on-site, with real-time price and availability sync.
• Advantage+ Shopping campaigns: Launched Meta’s Advantage+ Shopping (ASC) campaigns with the pixel and CAPI feeding audience signals. ASC was given 60% of budget with a broad targeting approach, while manual catalog campaigns retained 40% for category-specific creative testing.
• Creative by funnel: Developed a creative matrix by funnel stage — (1) Awareness (brand video, UGC styling shoots), (2) Consideration (product close-ups, fabric texture shots), (3) Conversion (price-anchor creatives, “Limited stock” urgency). Each ad set was served the appropriate creative mix based on placement and audience.
• Budget by margin: Allocated 65% of Meta budget to high-margin categories (unstitched fabrics) with 3x higher frequency caps, and 35% to pret wear with lower frequency but broader reach.

Campaign Metric (Pre-Rebuild)	Campaign Metric (Post-Rebuild)
Meta ROAS: 1.4x	Meta ROAS: 2.9x
Google Shopping ROAS: 1.8x	Google Shopping ROAS: 3.7x
Impression share (top categories): 18%	Impression share (top categories): 34%
Cost per click (CPC): PKR 12.40	Cost per click (CPC): PKR 8.60

Week 7 outcome: Combined ROAS across Meta and Google rose from 1.6x to 2.7x. The high-margin unstitched fabrics campaign hit 3.8x ROAS, while the sale items campaign stabilized at 2.1x (acceptable for clearance). CPA dropped from PKR 890 to PKR 620.

Phase 4 — Scale and Compound (Weeks 7–12)

The final phase focused on scaling the winning campaigns and building a self-sustaining optimization system. We implemented automated reporting, trained the internal team on feed management, and built playbooks for creative testing and budget allocation.

Scale actions:

• Budget reallocation: Shifted 30% of budget from low-performing campaigns (generic brand campaigns, broad lookalikes) to high-ROAS campaigns (unstitched fabrics Shopping, catalog sales DPAs). Total spend remained at PKR 1.2M/month, but efficiency improved dramatically.
• Creative testing framework: Built a 2-week creative testing cadence with statistical significance checks. Winning creatives (fabric texture videos, “Shop the look” carousels) were rolled out to 100% of relevant campaigns. Losing creatives (static plain backdrops, generic brand messages) were killed within 7 days.
• Feed automation: Implemented Shopify Flow to auto-update the Google Merchant Center feed when products were added/updated. Created a feed health dashboard in Looker Studio that monitored approval status, disapproval reasons, and missing attributes — alerting the team via Slack when feed health dropped below 90%.
• Internal training: Conducted a workshop for the marketing team on Shopify Flow basics, feed management, and campaign structure. Created a “Playbooks” Notion page with templates for campaign launches, creative tests, and feed diagnostics.

Phase 4 Initiative	Measured Impact
Budget reallocation to high-ROAS campaigns	ROAS from 2.7x to 3.3x (+22%)
Creative testing framework (4 winning variants)	+18% CTR on Meta, +12% on Google
Feed health automation (Shopify Flow + Slack alerts)	Feed approval sustained at 94%+
Internal training (2 sessions + playbooks)	3 internal feed/campaign optimizations launched post-engagement

Week 12 outcome: Combined ROAS reached 3.3x (up from 1.6x at baseline). Attributed monthly revenue from paid channels grew from PKR 1.92M to PKR 4.38M (+128%). CPA dropped from PKR 890 to PKR 515 (-42%), well under the PKR 450 target. The brand was now positioned to scale spend to PKR 1.8M/month while maintaining ROAS above 3.0x.

Final Results

At the 90-day mark (Day 84 post-launch), the engagement delivered measurable improvements across technical infrastructure, feed quality, and paid performance:

Metric	Before	After	Change
Return on Ad Spend (ROAS)	1.6x	3.3x	+106% lift
Cost per Acquisition (CPA)	PKR 890	PKR 515	-42%
Attributed monthly revenue (paid)	PKR 1.92M	PKR 4.38M	+128%
Product feed approval rate	68%	94%	+26 points
Largest Contentful Paint (LCP)	4.8s	1.9s	-60%
Checkout abandonment rate	72%	61%	-11 points
Google Rich Results eligibility	12%	78%	+66 points

The revenue impact transformed the unit economics. At 3.3x ROAS and PKR 515 CPA, the brand could profitably scale spend to PKR 1.8M/month, generating PKR 5.94M in monthly attributed revenue — a path to PKR 10M+ monthly revenue within 6 months. The Shopify Plus migration also positioned the brand for international expansion (multi-currency, multi-storefront) and headless commerce (Shopify as backend, custom frontend) when ready.

What Made This Work

Four success factors drove the outcome:

1. Platform-first, campaigns-second approach: We didn’t start by tweaking bids or creatives. The foundational issues — platform limits, slow page speeds, broken feeds — were constraining performance no matter how good the campaigns were. Fixing the platform first created the headroom for campaigns to actually scale.

2. Variant-level feed structure: Moving from a flat product feed to a variant-level feed unlocked query-specific matching in Google Shopping. A search for “blue lawn suit size 40” now matched the exact variant instead of a generic product page, which directly lifted CTR and conversion rate.

3. Margin-aware budget allocation: The brand had been spending blindly across all categories. By allocating budget to high-margin categories (unstitched fabrics at 3x margin) and treating sale items as a separate low-ROAS bucket, overall portfolio ROAS improved even when individual campaigns had different targets.

4. Local payment and checkout optimization: Surfacing JazzCash, EasyPaisa, and local bank options prominently in checkout, plus transparent shipping timelines by city, reduced checkout abandonment from 72% to 61%. This was a pure conversion win that required no ad spend — just better UX for the Pakistani customer.

What Teams Can Apply

Takeaways for Pakistani ecommerce teams:

1. Audit your feed health before scaling spend. If 30%+ of your products are disapproved in Google Merchant Center, you’re burning budget on wasted impressions. Run feed diagnostics, enrich attributes (color, size, material), and restructure to variant-level feeds before you increase spend.

2. Measure ROAS by margin, not just revenue. A 2.5x ROAS on a high-margin category is more valuable than a 3.5x ROAS on clearance items. Build margin-aware campaign structures and allocate budget where the unit economics actually support sustainable growth.

3. Platform debt compounds faster than you think. Shopify Basic was fine at 500 orders/month, but at 1,500+ orders/month, the app bloat, checkout limitations, and performance drag were silently killing ROAS. Migrate to Plus (or an equivalent scalable platform) before you hit the wall — not after.

4. Implement server-side tracking now. Client-side pixels alone are insufficient with privacy changes and ad-blockers. Server-side GTM and Facebook CAPI reduce data loss and improve attribution accuracy — this client saw a 15–20% uplift in attributed conversions post-implementation.

5. Localize checkout for Pakistani buyers. If your checkout assumes credit cards and international shipping, you’re losing 30–40% of interested buyers. Surfacing JazzCash, EasyPaisa, and local bank options, plus city-specific shipping timelines, is a pure conversion win that costs nothing but captures revenue you’re already paying to acquire.