Chapter 20: Phase 0 - Nexus MVP

Goal

Validate the AdPriority concept using Nexus Clothing as a real-world test case. Phase 0 is a manual implementation using Google Sheets as the supplemental feed transport, with no application code. The purpose is to prove that priority-based custom labels in Google Merchant Center produce measurable ROAS improvement in Performance Max campaigns before investing in SaaS development.

Status: IN PROGRESS - Supplemental feed validated (10/10 matched)

Steps Completed

Step 1: Analyze Nexus Catalog

Pulled live data from the Shopify Admin API for nexus-clothes.myshopify.com.

The catalog follows a hierarchical product type naming convention: {Gender}-{Department}-{SubCategory}-{Detail}. Examples include Men-Tops-T-Shirts, Men-Bottoms-Pants-Jeans, and Headwear-Baseball-Fitted. The 90 product types were consolidated into 20 category groups for manageable priority assignment (see Appendix C).

Step 2: Verify GMC ID Format

Exported the full Google Merchant Center product catalog (124,060 variants) and confirmed the product ID format.

CONFIRMED FORMAT
================

  shopify_US_{productId}_{variantId}

  Examples:
    shopify_US_8779355160808_46050142748904
    shopify_US_9128994570472_47260097118440
    shopify_US_9057367064808_47004004712680

All products in GMC use variant-level IDs. The Item Group ID field contains just the Shopify product ID (e.g., 8779355160808). Country code is US for all Nexus products.

Step 3: Confirm Custom Labels Available

Audited all five custom label slots in the GMC product export.

All five labels are effectively available. The 7 products with existing custom_label_0 values represent a negligible fraction of the catalog and are safe to overwrite.

Step 4: Create Sample Feed

Built a Google Sheet with 10 test products spanning multiple categories, priority levels, and seasons.

Feed Structure:

Columns:
  id                 - GMC product ID (shopify_US_{pid}_{vid})
  custom_label_0     - Priority score (priority-0 through priority-5)
  custom_label_1     - Season (winter)
  custom_label_2     - Category group (e.g., headwear-caps, outerwear-heavy)
  custom_label_3     - Product status (in-stock, low-inventory, dead-stock)
  custom_label_4     - Brand tier (name-brand, off-brand)

Sample Products Tested:

Step 5: Upload to GMC

Connected the Google Sheet as a supplemental feed in Google Merchant Center.

Process:

1. Shared the Google Sheet publicly (Viewer access)
2. In GMC: Products > Feeds > Add supplemental feed > Google Sheets
3. Linked the feed to all three primary data sources:
   • Content API - US, English
   • Content API - Local, US
   • Local Feed Partnership
4. Triggered a manual update

Results:

VALIDATION RESULT
=================

  Feed Upload        -->  10 rows submitted
  ID Matching        -->  10/10 matched (100%)
  Label Application  -->  All 5 custom labels applied
  Errors             -->  0
  Warnings           -->  0
  Status             -->  VALIDATED

Remaining Steps

Step 6: Expand to Full Active Catalog

Scale the supplemental feed from 10 test products to the full active catalog.

Scope:

Process:

1. Export all active products from Shopify API (paginated, 250/page, ~10 pages)
2. For each product, enumerate all variants
3. Apply category mapping rules (20 groups from category-mapping.md)
4. Apply seasonal priorities (Winter season as of February 2026)
5. Apply tag modifiers (NAME BRAND +1, in-stock +1, Sale -1, DEAD50 override to 0, archived override to 0, warning_inv_1 -1)
6. Apply new arrival boost (products created within 14 days get priority 5)
7. Generate GMC ID for each variant: shopify_US_{productId}_{variantId}
8. Write all rows to Google Sheet via Sheets API or manual paste
9. Verify GMC processes the full feed without errors

Expected Distribution (Winter, estimated):

PROJECTED PRIORITY DISTRIBUTION
================================

  Priority 5 (Push Hard):     ~300  (12%)  Puffer jackets, hoodies, new arrivals
  Priority 4 (Strong):        ~500  (21%)  Jeans, name brand items, joggers
  Priority 3 (Normal):        ~800  (33%)  Caps, basic t-shirts, year-round items
  Priority 2 (Low):           ~400  (16%)  Underwear, socks, off-season light items
  Priority 1 (Minimal):       ~200   (8%)  End-of-season, slow movers
  Priority 0 (Exclude):       ~225  (10%)  Dead stock, archived, out of stock
                              -----
  Total Active:              2,425

Step 7: Set Up PMAX Campaigns by Priority Tier

Restructure existing Google Ads Performance Max campaigns to segment products by the custom_label_0 priority score.

Proposed Campaign Structure:

PMAX CAMPAIGN STRUCTURE
=======================

  Campaign 1: "PMAX - High Priority (5)"
    Listing group filter: custom_label_0 = "priority-5"
    Daily budget: $50
    Target ROAS: 3.0x
    Products: ~300

  Campaign 2: "PMAX - Strong (4)"
    Listing group filter: custom_label_0 = "priority-4"
    Daily budget: $35
    Target ROAS: 4.0x
    Products: ~500

  Campaign 3: "PMAX - Normal (3)"
    Listing group filter: custom_label_0 = "priority-3"
    Daily budget: $20
    Target ROAS: 5.0x
    Products: ~800

  Campaign 4: "PMAX - Low (1-2)"
    Listing group filter: custom_label_0 IN ("priority-1", "priority-2")
    Daily budget: $10
    Target ROAS: 6.0x
    Products: ~600

  Priority 0 products: Excluded (not in any campaign)

Budget Allocation (daily):

Step 8: Monitor ROAS for 30 Days

Track campaign performance daily for 30 days to establish whether the priority-segmented structure outperforms the previous flat structure.

Baseline Metrics (to capture before restructuring):

Success Criteria:

Tracking Spreadsheet:

DATE        | CAMPAIGN     | SPEND  | REVENUE | ROAS  | CONV | CPA
------------|------------- |--------|---------|-------|------|------
2026-02-15  | High (5)     | $48.20 | $241.00 | 5.0x  |  24  | $2.01
2026-02-15  | Strong (4)   | $33.10 | $132.40 | 4.0x  |  17  | $1.95
2026-02-15  | Normal (3)   | $19.50 |  $58.50 | 3.0x  |  10  | $1.95
2026-02-15  | Low (1-2)    |  $9.80 |  $24.50 | 2.5x  |   5  | $1.96
...

Step 9: Document Results

Produce a comprehensive results document covering:

1. Methodology: How products were scored, how campaigns were structured
2. Baseline vs. Result: Before/after comparison with statistical significance
3. Category Analysis: Which category groups benefited most from prioritization
4. Seasonal Validation: Whether Winter priorities matched actual demand
5. Lessons Learned: What to adjust in the SaaS version
6. Process Documentation: Step-by-step replication guide for Phase 1

Tools and Resources

Success Metric

The single success metric for Phase 0 is:

Measurable ROAS improvement versus the flat (unsegmented) campaign structure, sustained over a 30-day monitoring period.

If ROAS improves by 20% or more with the priority-segmented approach, the concept is validated and Phase 1 (SaaS development) proceeds. If ROAS does not improve, the category mappings and seasonal priorities are revised before retesting.

Timeline

PHASE 0 TIMELINE
================

  Week 1 (Feb 3-9):
    [DONE] Analyze catalog (5,582 products, 90 types)
    [DONE] Verify GMC ID format
    [DONE] Confirm custom labels available
    [DONE] Create sample feed (10 products)
    [DONE] Upload and validate sample feed (10/10 matched)

  Week 2 (Feb 10-16):
    [TODO] Expand feed to full active catalog (~2,425 products)
    [TODO] Capture baseline PMAX metrics
    [TODO] Restructure PMAX campaigns by priority tier

  Week 3-6 (Feb 17 - Mar 16):
    [TODO] Monitor ROAS daily for 30 days
    [TODO] Weekly progress checks
    [TODO] Adjust priorities if data warrants

  Week 7 (Mar 17-23):
    [TODO] Compile results document
    [TODO] Go/no-go decision for Phase 1