Shopify Admin Variant Performance Report

Purpose

Goes beyond product-level revenue by ranking every individual variant (size, color, option combination) on revenue, units sold, and refund rate, then joining against live inventory levels. Reveals which specific SKUs are driving the business and which are tying up capital on the shelf. Read-only — no mutations are executed.

Prerequisites

• Authenticated Shopify CLI session: shopify auth login --store <domain>
• API scopes: read_orders, read_products (validator-confirmed: orders query traverses variant→product graph)

Parameters

Workflow Steps

1. OPERATION: orders — query Inputs: first: 250, query: "created_at:>='<date_range_start>' created_at:<='<date_range_end>'", pagination cursor; select lineItems with variant { id, sku, title, selectedOptions }, quantity, originalTotalSet; and refunds.refundLineItems with variant id and subtotalSet Expected output: All orders in range; paginate until hasNextPage: false; aggregate in-memory per variant.id: units sold, gross revenue, refunded units, refunded amount, refund rate

2. OPERATION: productVariants — query Inputs: List of variant IDs collected in step 1, first: 250, pagination cursor; select id, sku, title, selectedOptions, inventoryQuantity, product { id, title }, price Expected output: Current inventory levels and metadata for each sold variant; joined with step-1 aggregates; variants present in inventory but with zero sales are flagged as dead stock candidates

3. In-memory computation: Sort merged dataset by sort_by metric; compute revenue-per-inventory-unit ratio (net revenue ÷ inventory quantity) to highlight variants earning little relative to shelf space; split output into top-N performers and bottom-N by the same metric

GraphQL Operations

graphqlCopy code
# orders:query (variant line items + refunds) — validated against api_version 2025-01
query OrdersForVariantPerformance($first: Int!, $after: String, $query: String) {
  orders(first: $first, after: $after, query: $query) {
    edges {
      node {
        id
        createdAt
        lineItems(first: 50) {
          edges {
            node {
              quantity
              originalTotalSet {
                shopMoney { amount currencyCode }
              }
              variant {
                id
                sku
                title
                selectedOptions { name value }
                product { id title }
              }
            }
          }
        }
        refunds {
          refundLineItems(first: 50) {
            edges {
              node {
                quantity
                subtotalSet {
                  shopMoney { amount currencyCode }
                }
                lineItem {
                  variant { id }
                }
              }
            }
          }
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

graphqlCopy code
# productVariants:query (inventory snapshot) — validated against api_version 2025-01
query VariantInventorySnapshot($first: Int!, $after: String, $query: String) {
  productVariants(first: $first, after: $after, query: $query) {
    edges {
      node {
        id
        sku
        title
        price
        selectedOptions { name value }
        inventoryQuantity
        product {
          id
          title
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

Session Tracking

Claude MUST emit the following output at each stage. This is mandatory.

On start, emit:

Copy code
╔══════════════════════════════════════════════╗
║  SKILL: variant-performance-report           ║
║  Store: <store domain>                       ║
║  Started: <YYYY-MM-DD HH:MM UTC>             ║
╚══════════════════════════════════════════════╝

After each step, emit:

Copy code
[N/TOTAL] <QUERY|MUTATION>  <OperationName>
          → Params: <brief summary of key inputs>
          → Result: <count or outcome>

On completion, emit:

For format: human (default):

Copy code
══════════════════════════════════════════════
OUTCOME SUMMARY
  Orders processed:     <n>
  Variants analysed:    <n>
  Date range:           <start> to <end>
  Sort by:              <metric>
  Errors:               0
  Output:               variant_performance_<date>.csv
══════════════════════════════════════════════

For format: json, emit:

jsonCopy code
{
  "skill": "variant-performance-report",
  "store": "<domain>",
  "started_at": "<ISO8601>",
  "completed_at": "<ISO8601>",
  "dry_run": false,
  "steps": [
    { "step": 1, "operation": "OrdersForVariantPerformance", "type": "query", "params_summary": "<date_range_start> to <date_range_end>", "result_summary": "<n> orders, <n> variants aggregated", "skipped": false },
    { "step": 2, "operation": "VariantInventorySnapshot", "type": "query", "params_summary": "<n> variant IDs", "result_summary": "<n> variants with inventory data", "skipped": false }
  ],
  "outcome": {
    "orders_processed": 0,
    "variants_analysed": 0,
    "date_range_start": "<date_range_start>",
    "date_range_end": "<date_range_end>",
    "sort_by": "revenue",
    "top_performers": [],
    "dead_weight": [],
    "errors": 0,
    "output_file": "variant_performance_<date>.csv"
  }
}

Output Format

CSV file variant_performance_<YYYY-MM-DD>.csv with one row per variant:

For format: human, two ranked tables are printed inline:

1. Top performers — top top_n variants by sort_by metric
2. Dead weight — bottom top_n variants by revenue_per_inventory_unit (≥ min_units sold, inventory > 0)

Error Handling

Best Practices

1. Run with a 30–90 day window first. Very wide windows produce large pagination chains and slow down step 1 significantly.
2. The revenue_per_inventory_unit column is the sharpest signal for dead weight — a high inventory count with near-zero revenue is a clear markdown candidate.
3. High refund_rate_pct on a specific size or color often points to a fit or quality issue — investigate before reordering that option.
4. Use min_units: 5 to filter out statistical noise from variants with very few sales before making merchandising decisions.
5. Pair with dead-stock-identifier for a broader view of inventory health beyond the sales period captured here.