Shopify Admin Abandoned Cart Recovery

Purpose

Identifies customers who started checkout but did not complete their purchase, generates a unique discount code for each one, and tags them in Shopify so they can be targeted in follow-up campaigns. This skill handles the Shopify-native data layer (querying, discounts, tagging); sending the actual email requires an external tool.

Prerequisites

• Authenticated Shopify CLI session: shopify auth login --store <domain>
• API scopes: read_checkouts, write_price_rules, write_discount_codes, write_customers

Parameters

Safety

⚠️ Steps 2 and 3 execute mutations (discount code creation, customer tagging). Discount codes created via discountCodeBulkCreate cannot be bulk-deleted via the Admin API — they must be removed individually or via the price rule. Run with dry_run: true to verify eligible customer count before committing.

Workflow Steps

1. OPERATION: abandonedCheckouts — query Inputs: first: 250, query: "created_at:>='<NOW - days_back days>'", pagination cursor Expected output: List of checkout objects with email, totalPrice, createdAt, lineItems; paginate until hasNextPage: false

2. OPERATION: discountCodeBulkCreate — mutation Inputs: For each eligible customer email: a unique code {code_prefix}-{UUID[:8].toUpperCase()}, percentage discount discount_pct, usage limit 1, expiry 30 days Expected output: Confirmation of code creation with code and priceRule.id; collect userErrors

3. OPERATION: tagsAdd — mutation Inputs: Customer id (from checkout email lookup), tag string from tag parameter Expected output: Updated customer tags; collect userErrors

GraphQL Operations

# abandonedCheckouts:query — validated against api_version 2025-04
query AbandonedCheckouts($first: Int!, $after: String, $query: String) {
  abandonedCheckouts(first: $first, after: $after, query: $query) {
    edges {
      node {
        id
        customer {
          defaultEmailAddress {
            emailAddress
          }
        }
        totalPriceSet {
          shopMoney {
            amount
            currencyCode
          }
        }
        createdAt
        lineItems {
          edges {
            node {
              title
              quantity
              variant {
                price
              }
            }
          }
        }
      }
    }
    pageInfo {
      hasNextPage
      endCursor
    }
  }
}

# discountCodeBulkCreate:mutation — validated against api_version 2025-01
mutation DiscountCodeBulkCreate($priceRuleId: ID!, $codes: [DiscountCodeInput!]!) {
  discountCodeBulkCreate(priceRuleId: $priceRuleId, codes: $codes) {
    bulkCreations {
      id
      done
    }
    userErrors {
      field
      message
    }
  }
}

# tagsAdd:mutation — validated against api_version 2025-01
mutation TagsAdd($id: ID!, $tags: [String!]!) {
  tagsAdd(id: $id, tags: $tags) {
    node {
      id
    }
    userErrors {
      field
      message
    }
  }
}

Session Tracking

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

On start, emit:

╔══════════════════════════════════════════════╗
║  SKILL: Abandoned Cart Recovery              ║
║  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>

If dry_run: true, prefix every mutation step with [DRY RUN] and do not execute it.

On completion, emit:

For format: human (default):

══════════════════════════════════════════════
OUTCOME SUMMARY
  Checkouts found:     <n>
  Eligible customers:  <n>
  Codes created:       <n>
  Customers tagged:    <n>
  Errors:              <n>
  Output:              recovery_list_<date>.csv
══════════════════════════════════════════════

For format: json, emit:

{
  "skill": "abandoned-cart-recovery",
  "store": "<domain>",
  "started_at": "<ISO8601>",
  "completed_at": "<ISO8601>",
  "dry_run": false,
  "steps": [
    { "step": 1, "operation": "AbandonedCheckouts", "type": "query", "params_summary": "last 7 days, cart >= $0", "result_summary": "<n> checkouts", "skipped": false },
    { "step": 2, "operation": "DiscountCodeBulkCreate", "type": "mutation", "params_summary": "<n> codes, 10% off, prefix RECOVER", "result_summary": "<n> created", "skipped": false },
    { "step": 3, "operation": "TagsAdd", "type": "mutation", "params_summary": "tag: cart-recovery", "result_summary": "<n> customers tagged", "skipped": false }
  ],
  "outcome": {
    "checkouts_found": 0,
    "eligible_customers": 0,
    "codes_created": 0,
    "customers_tagged": 0,
    "errors": 0,
    "output_file": "recovery_list_<date>.csv"
  }
}

Output Format

CSV file recovery_list_<YYYY-MM-DD>.csv with columns: customer_email, cart_total, abandoned_at, discount_code, discount_pct, tag_applied

Error Handling

Best Practices

• Run with dry_run: true first and confirm eligible customer count — discount codes cannot be bulk-deleted via API and will persist in your price rules until manually removed.
• Set min_cart_value to focus recovery effort on higher-value carts (e.g., 50 for $50+). Small carts often have lower recovery intent and may not justify a discount.
• Use a unique tag per run (e.g., cart-recovery-2026-04-11) so you can identify exactly which customers were targeted in each campaign and avoid re-targeting them in subsequent runs.
• The code_prefix should be meaningful to operators reviewing discount codes in Shopify Admin (e.g., CART10 for a 10% cart recovery discount).
• This skill produces the data layer. Use Shopify Email or another email tool to send the generated codes to customers.