Discount Ab Analysis

Compare redemption rates and revenue performance across two or more discount codes over a specified date range.

shopify-admin-discount-ab-analysis


Purpose

Compares how different discount codes perform against each other by redemption count and revenue generated. Useful for A/B testing promotional offers without a dedicated analytics app — provide two or more codes and a date range, and the skill queries Shopify for discount metadata and order revenue, then produces a side-by-side comparison table. Read-only: no mutations are executed.


Prerequisites

  • Authenticated Shopify CLI session: shopify auth login --store
  • API scopes: read_discounts, read_orders

    • Parameters


    ParameterTypeRequiredDefaultDescription
    storestringyes—Store domain (e.g., mystore.myshopify.com)
    formatstringnohumanOutput format: human or json
    dry_runboolnofalsePreview operations without executing mutations
    discount_codesarrayyes—Array of 2 or more discount code strings to compare (e.g., ["SAVE10", "WELCOME15"])
    date_range_startstringyes—Start date in ISO 8601 (e.g., 2025-01-01)
    date_range_endstringyes—End date in ISO 8601 (e.g., 2025-01-31)

    Workflow Steps


  • OPERATION: discountNodes — query

    • Inputs: first: 50, query: "code:" (one query per code in discount_codes)

    Expected output: Discount metadata: title, code strings, asyncUsageCount, status, startsAt, endsAt per code


  • OPERATION: orders — query (one paginated query per discount code)

    • Inputs: first: 250, query: "discount_code: created_at:>='' created_at:<=''", pagination cursor

    Expected output: Orders containing the discount code with totalPriceSet; paginate until hasNextPage: false; aggregate: count, sum revenue, compute avg order value


    GraphQL Operations


    # discountNodes:query — validated against api_version 2025-01
query DiscountNodes($first: Int!, $query: String) {
  discountNodes(first: $first, query: $query) {
    edges {
      node {
        id
        discount {
          ... on DiscountCodeBasic {
            title
            codes(first: 10) {
              edges {
                node {
                  code
                  asyncUsageCount
                }
              }
            }
            usageLimit
            status
            startsAt
            endsAt
          }
          ... on DiscountCodeBxgy {
            title
            codes(first: 10) {
              edges {
                node {
                  code
                  asyncUsageCount
                }
              }
            }
            status
          }
          ... on DiscountCodeFreeShipping {
            title
            codes(first: 10) {
              edges {
                node {
                  code
                  asyncUsageCount
                }
              }
            }
            status
          }
        }
      }
    }
  }
}


    # orders:query (by discount code) — validated against api_version 2025-01
query OrdersByDiscountCode($first: Int!, $after: String, $query: String) {
  orders(first: $first, after: $after, query: $query) {
    edges {
      node {
        id
        createdAt
        totalPriceSet {
          shopMoney { amount currencyCode }
        }
        discountCodes
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

    Session Tracking


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


    On start, emit:

    ╔══════════════════════════════════════════════╗
║  SKILL: discount-ab-analysis                 ║
║  Store: <store domain>                       ║
║  Started: <YYYY-MM-DD HH:MM UTC>             ║
╚══════════════════════════════════════════════╝

    After each step, emit:

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

    On completion, emit:


    For format: human (default):

    ══════════════════════════════════════════════
OUTCOME SUMMARY
  Codes analyzed:   <n>
  Date range:       <start> to <end>
  Errors:           0
  Output:           none
══════════════════════════════════════════════

    For format: json, emit:

    {
  "skill": "discount-ab-analysis",
  "store": "<domain>",
  "started_at": "<ISO8601>",
  "completed_at": "<ISO8601>",
  "dry_run": false,
  "steps": [
    { "step": 1, "operation": "DiscountNodes", "type": "query", "params_summary": "<n> codes queried", "result_summary": "<n> discount nodes found", "skipped": false },
    { "step": 2, "operation": "OrdersByDiscountCode", "type": "query", "params_summary": "date range <start> to <end>", "result_summary": "<n> orders aggregated", "skipped": false }
  ],
  "outcome": {
    "codes_analyzed": 0,
    "date_range_start": "<start>",
    "date_range_end": "<end>",
    "results": [
      {
        "code": "SAVE10",
        "async_usage_count": 0,
        "orders_in_range": 0,
        "total_revenue": "0.00",
        "avg_order_value": "0.00",
        "revenue_per_use": "0.00"
      }
    ],
    "errors": 0,
    "output_file": null
  }
}

    Output Format


    A comparison table per code (displayed inline):


    CodeUses (asyncUsageCount)Orders in RangeTotal RevenueAvg Order ValueRevenue per Use
    SAVE10...............
    WELCOME15...............

    For format: json, the results array contains one object per code with keys: code, async_usage_count, orders_in_range, total_revenue, avg_order_value, revenue_per_use.


    Error Handling

    ErrorCauseRecovery
    Discount code not foundCode doesn't exist or was deletedVerify code in Shopify admin
    No orders returned for a codeNo orders used this code in the date rangeWiden date range or verify code was active
    discount_codes has fewer than 2 entriesCan't do A/B with 1 codeProvide at least 2 codes
    Rate limit (429)Too many paginated orders queriesWait and retry; reduce date range

    Best Practices

  • asyncUsageCount is the lifetime usage count from the discount object — orders_in_range is what was redeemed in your date window. Both are reported for full context.
  • For codes with high usage, the orders query will paginate — larger date ranges may produce many API calls. Consider narrowing the date range for faster results.
  • Revenue per use is the best signal for comparing codes with different usage volumes.
  • Run this analysis at the end of a campaign period before deciding which discount strategy to repeat.
  • If asyncUsageCount is 0 for a code, check that the code was active during the date range and correctly applied at checkout.