Find Wasted Meta Ad Spend by Placement, Region, Age and Hour with Claude

The workflow JSON

Copy or download the full n8n JSON below. Paste it into a new n8n workflow, add your credentials, activate. Full import guide →

{
  "id": "mvxVBIvdjGEVWgGX",
  "meta": {
    "templateCredsSetupCompleted": true
  },
  "name": "Find wasted Meta ad spend by placement, region, age and hour with AI",
  "tags": [],
  "nodes": [
    {
      "id": "3f08678a-3807-4bad-b04c-94fb90890617",
      "name": "Overview",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        32176,
        5104
      ],
      "parameters": {
        "width": 732,
        "height": 632,
        "content": "## Meta Ads Wasted-Spend Finder\n\nFinds the delivery slices of your Meta ad account that waste budget (by placement, age/gender, hour, and region) and gives you a prioritized, AI-ranked kill list with a recommended action for each. Works for lead-gen and ecommerce.\n\n\n### How it works\n- Pulls last-30-day account insights plus four delivery breakdowns from the Meta Marketing API.\n- Sums your configured result events (leads, registrations, purchases) per slice and computes the account cost-per-result benchmark.\n- Auto-skips any breakdown Meta does not attribute conversions to, so you never get false \"zero result\" waste.\n- Flags slices spending over a floor with zero results, or a cost-per-result far above benchmark, and scores the dollars recoverable.\n- Claude ranks the kill list and recommends pause, exclude, or reallocate.\n\n\n### Setup\n1. Add to your .env: META_ACCESS_TOKEN, META_AD_ACCOUNT_ID, META_API_VERSION, ANTHROPIC_API_KEY.\n2. Run with the manual trigger, or attach a Schedule trigger for a daily report.\n\n### Customization\nIn the Config node, set RESULT_ACTIONS to the events you optimize for, and tune SPEND_FLOOR and CPR_MULTIPLE. Swap the stdout report for a Slack or email node to auto-deliver.\n\nBuilt by https://nocode.expert - done-for-you automation & tracking. "
      },
      "typeVersion": 1
    },
    {
      "id": "289d952b-7405-4fa3-a85d-bf2f35b427d8",
      "name": "Section: Fetch & benchmark",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        33168,
        5200
      ],
      "parameters": {
        "color": 7,
        "width": 720,
        "height": 704,
        "content": "## 1. Fetch & benchmark\nPull Meta spend + results per breakdown; skip breakdowns with no conversion attribution."
      },
      "typeVersion": 1
    },
    {
      "id": "fd9ec440-1dc9-4982-805d-495b66f1d61d",
      "name": "Section: Detect waste",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        33984,
        5200
      ],
      "parameters": {
        "color": 7,
        "width": 800,
        "height": 704,
        "content": "## 2. Detect & rank waste\nFlag slices below benchmark, then Claude ranks the kill list."
      },
      "typeVersion": 1
    },
    {
      "id": "21fa710c-fdb7-4583-913a-4feb21e9a85e",
      "name": "Section: Report",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        34880,
        5200
      ],
      "parameters": {
        "color": 7,
        "width": 1740,
        "height": 704,
        "content": "## 3. Report\nPrint the prioritized report (swap for Slack/email)."
      },
      "typeVersion": 1
    },
    {
      "id": "7d3def97-da0c-437e-9eb2-3004633e820c",
      "name": "Run manually",
      "type": "n8n-nodes-base.manualTrigger",
      "position": [
        32992,
        5456
      ],
      "parameters": {},
      "typeVersion": 1
    },
    {
      "id": "bdc63539-6962-4cf3-84c7-4bd51f2b324f",
      "name": "Flag underperforming slices vs benchmark",
      "type": "n8n-nodes-base.code",
      "position": [
        34944,
        5456
      ],
      "parameters": {
        "jsCode": "// Flag slices wasting budget vs the account's own cost-per-result benchmark:\n// zero results with real spend, or a cost-per-result far above benchmark.\n// \"At risk\" = dollars spent above what benchmark efficiency would have cost.\nconst { slices, skipped, benchmarkCPR, acctSpend, acctResults, cfg } = $json;\nconst FLOOR = cfg.SPEND_FLOOR, MULT = cfg.CPR_MULTIPLE;\n\nconst wasted = (slices || [])\n  .filter((s) => s.spend >= FLOOR && (s.results === 0 || (benchmarkCPR && s.cpr > benchmarkCPR * MULT)))\n  .map((s) => ({\n    ...s,\n    waste_type: s.results === 0 ? 'zero_results' : 'high_cost_per_result',\n    at_risk: s.results === 0 ? s.spend : +Math.max(0, s.spend - s.results * benchmarkCPR).toFixed(2),\n  }))\n  .sort((a, b) => b.at_risk - a.at_risk)\n  .slice(0, 20);\n\nconst total_at_risk = +wasted.reduce((n, s) => n + s.at_risk, 0).toFixed(2);\nreturn [{ json: { wasted, total_at_risk, benchmarkCPR, acctSpend, acctResults, skipped, floor: FLOOR, cpr_multiple: MULT, considered: (slices || []).length } }];"
      },
      "typeVersion": 2
    },
    {
      "id": "653dc56a-87e2-4df2-b7c4-41a6b3f7ebe0",
      "name": "Build AI prioritization prompt",
      "type": "n8n-nodes-base.code",
      "position": [
        35712,
        5248
      ],
      "parameters": {
        "jsCode": "// Ask Claude to rank the kill list and recommend a concrete action per slice.\nconst { wasted, total_at_risk, benchmarkCPR } = $json;\nconst system = 'You are a senior paid-media analyst. You are given underperforming delivery slices from a Meta ad account, judged against the account cost-per-result benchmark. Rank them by dollars recoverable, and for each give a one-line rationale and one concrete action: pause, exclude, or reallocate. Be specific and terse. Return ONLY valid JSON, no prose.';\nconst user = 'Account cost-per-result benchmark: $' + benchmarkCPR + '. Total at risk: $' + total_at_risk + '.\\n\\nSlices:\\n' + JSON.stringify(wasted) + '\\n\\nReturn JSON exactly: {\"summary\":\"one sentence\",\"ranked\":[{\"slice\":\"\",\"dimension\":\"\",\"at_risk\":0,\"action\":\"pause|exclude|reallocate\",\"why\":\"\"}]}';\nconst body = { model: 'claude-haiku-4-5', max_tokens: 1500, system, messages: [{ role: 'user', content: user }] };\nreturn [{ json: { body, _ctx: $json } }];"
      },
      "typeVersion": 2
    },
    {
      "id": "b250ddda-aea3-484a-b2f0-4e0a3d5133a2",
      "name": "Rank the kill list with Claude AI",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        35936,
        5248
      ],
      "parameters": {
        "url": "https://api.anthropic.com/v1/messages",
        "method": "POST",
        "options": {},
        "jsonBody": "={{ $json.body }}",
        "sendBody": true,
        "sendHeaders": true,
        "specifyBody": "json",
        "headerParameters": {
          "parameters": [
            {
              "name": "x-api-key",
              "value": "={{ $env.ANTHROPIC_API_KEY }}"
            },
            {
              "name": "anthropic-version",
              "value": "2023-06-01"
            },
            {
              "name": "content-type",
              "value": "application/json"
            }
          ]
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "6f900e46-9952-469b-bda6-1faaf8037c71",
      "name": "Print prioritized wasted-spend report",
      "type": "n8n-nodes-base.code",
      "position": [
        36256,
        5264
      ],
      "parameters": {
        "jsCode": "// Print the ranked wasted-spend report to stdout. Extract the JSON object by\n// braces so model code-fences never break parsing.\nconst ctx = $node['Flag underperforming slices vs benchmark'].json;\nconst raw = ($json.content && $json.content[0] && $json.content[0].text) || '{}';\nconst s = raw.indexOf('{'), e = raw.lastIndexOf('}');\nlet parsed = { summary: '(could not parse model output)', ranked: [] };\ntry { parsed = JSON.parse(raw.slice(s, e + 1)); } catch (err) {}\n\nconst out = [];\nout.push('');\nout.push('==========================================================');\nout.push('  META ADS WASTED-SPEND FINDER  |  nocode.expert');\nout.push('==========================================================');\nout.push('Account cost-per-result benchmark: $' + ctx.benchmarkCPR + '  (' + ctx.acctResults.toLocaleString() + ' results on $' + ctx.acctSpend.toLocaleString() + ')');\nif (ctx.skipped && ctx.skipped.length) {\n  out.push('Skipped (no conversion attribution): ' + ctx.skipped.map((x) => x.dimension).join(', '));\n}\nout.push('Total at risk: $' + (ctx.total_at_risk || 0).toLocaleString());\nout.push('');\nout.push(parsed.summary || '');\nout.push('');\n(parsed.ranked || []).forEach((r, i) => {\n  out.push(String(i + 1).padStart(2) + '. [$' + (r.at_risk || 0).toLocaleString() + ']  ' + r.dimension + ' :: ' + r.slice);\n  out.push('     -> ' + String(r.action || '').toUpperCase() + ': ' + (r.why || ''));\n});\nout.push('==========================================================');\nconsole.log(out.join('\\n'));\nreturn [{ json: parsed }];"
      },
      "typeVersion": 2
    },
    {
      "id": "5d1a11d8-5aa4-4001-93e8-8a03d544be9a",
      "name": "Fetch account totals (Meta)",
      "type": "n8n-nodes-base.facebookGraphApi",
      "position": [
        33472,
        5456
      ],
      "parameters": {
        "edge": "insights",
        "node": "={{ ($env.META_AD_ACCOUNT_ID || '').trim() }}",
        "options": {
          "queryParameters": {
            "parameter": [
              {
                "name": "level",
                "value": "account"
              },
              {
                "name": "date_preset",
                "value": "={{ $json.DATE_PRESET }}"
              },
              {
                "name": "fields",
                "value": "spend,impressions,clicks,actions"
              },
              {
                "name": "limit",
                "value": "500"
              }
            ]
          }
        },
        "graphApiVersion": "v21.0"
      },
      "typeVersion": 1
    },
    {
      "id": "337204af-cffe-4ec8-9eeb-162cbac65659",
      "name": "Compute benchmark",
      "type": "n8n-nodes-base.code",
      "position": [
        33680,
        5456
      ],
      "parameters": {
        "jsCode": "// Sum configured result events on the account total and compute the\n// cost-per-result benchmark. Pure data transform, no API calls.\nconst cfg = $('Set config: result events, floor').first().json;\nconst RES = new Set(cfg.RESULT_ACTIONS);\nconst resultsOf = (a) => Array.isArray(a)\n  ? a.filter(x => RES.has(x.action_type)).reduce((n,x)=>n+(Number(x.value)||0),0) : 0;\nconst acct = ($json.data && $json.data[0]) || {};\nconst acctSpend = Number(acct.spend) || 0;\nconst acctResults = resultsOf(acct.actions);\nconst benchmarkCPR = acctResults > 0 ? +(acctSpend/acctResults).toFixed(2) : null;\nreturn [{ json: { acctSpend:+acctSpend.toFixed(2), acctResults, benchmarkCPR } }];"
      },
      "typeVersion": 2
    },
    {
      "id": "4dad15a5-2d91-4bef-b683-41d2045001e5",
      "name": "List delivery breakdowns",
      "type": "n8n-nodes-base.code",
      "position": [
        34128,
        5456
      ],
      "parameters": {
        "jsCode": "// Emit one item per delivery breakdown. The next Facebook node then runs once\n// per item, so n8n loops over the breakdowns for you. Pure data transform.\nconst cfg = $('Set config: result events, floor').first().json;\nconst BREAKDOWNS = [\n  { dim: 'placement',  breakdowns: 'publisher_platform,platform_position' },\n  { dim: 'region',     breakdowns: 'region' },\n  { dim: 'age_gender', breakdowns: 'age,gender' },\n  { dim: 'hour',       breakdowns: 'hourly_stats_aggregated_by_advertiser_time_zone' },\n];\nreturn BREAKDOWNS.map(b => ({ json: { ...b, DATE_PRESET: cfg.DATE_PRESET } }));"
      },
      "typeVersion": 2
    },
    {
      "id": "ac9c364c-569d-4c6f-8218-e209deac11b9",
      "name": "Fetch breakdown insights (Meta)",
      "type": "n8n-nodes-base.facebookGraphApi",
      "position": [
        34336,
        5456
      ],
      "parameters": {
        "edge": "insights",
        "node": "={{ ($env.META_AD_ACCOUNT_ID || '').trim() }}",
        "options": {
          "queryParameters": {
            "parameter": [
              {
                "name": "level",
                "value": "account"
              },
              {
                "name": "date_preset",
                "value": "={{ $json.DATE_PRESET }}"
              },
              {
                "name": "fields",
                "value": "spend,impressions,clicks,actions"
              },
              {
                "name": "breakdowns",
                "value": "={{ $json.breakdowns }}"
              },
              {
                "name": "limit",
                "value": "500"
              }
            ]
          }
        },
        "graphApiVersion": "v21.0"
      },
      "typeVersion": 1
    },
    {
      "id": "9b21e500-ff29-4797-b55f-0299d3225a81",
      "name": "Flatten & score slices",
      "type": "n8n-nodes-base.code",
      "position": [
        34576,
        5456
      ],
      "parameters": {
        "jsCode": "// Combine the four breakdown responses into per-slice rows, apply the attribution\n// gate (skip a breakdown that loses most results), attach cost-per-result.\n// Pure data transform over the Facebook node responses.\nconst cfg = $('Set config: result events, floor').first().json;\nconst bench = $('Compute benchmark').first().json;\nconst reqs = $('List delivery breakdowns').all();\nconst RES = new Set(cfg.RESULT_ACTIONS);\nconst resultsOf = (a) => Array.isArray(a)\n  ? a.filter(x => RES.has(x.action_type)).reduce((n,x)=>n+(Number(x.value)||0),0) : 0;\nconst LABEL = {\n  placement:  (r) => `${r.publisher_platform} / ${r.platform_position}`,\n  region:     (r) => r.region,\n  age_gender: (r) => `${r.age} ${r.gender}`,\n  hour:       (r) => (r.hourly_stats_aggregated_by_advertiser_time_zone || '').slice(0,5),\n};\nconst responses = $input.all();\nconst slices = [], skipped = [];\nresponses.forEach((item, i) => {\n  const dim  = reqs[i].json.dim;\n  const rows = item.json.data || [];\n  const bResults = rows.reduce((n,r)=>n+resultsOf(r.actions),0);\n  if (bench.acctResults > 0 && bResults < bench.acctResults * cfg.MIN_ATTRIBUTION) {\n    skipped.push({ dimension: dim, captured: bResults, of: bench.acctResults }); return;\n  }\n  for (const r of rows) {\n    const spend = Number(r.spend)||0, results = resultsOf(r.actions);\n    slices.push({ dimension: dim, slice: LABEL[dim](r), spend:+spend.toFixed(2), results, cpr: results>0 ? +(spend/results).toFixed(2) : null });\n  }\n});\nreturn [{ json: { slices, skipped, benchmarkCPR: bench.benchmarkCPR, acctSpend: bench.acctSpend, acctResults: bench.acctResults, cfg } }];"
      },
      "typeVersion": 2
    },
    {
      "id": "326e79dc-c133-4316-b0c2-8de00d9d3934",
      "name": "Any budget at risk?",
      "type": "n8n-nodes-base.if",
      "position": [
        35360,
        5456
      ],
      "parameters": {
        "options": {},
        "conditions": {
          "options": {
            "version": 2,
            "leftValue": "",
            "caseSensitive": true,
            "typeValidation": "strict"
          },
          "combinator": "and",
          "conditions": [
            {
              "id": "cond-atrisk",
              "operator": {
                "type": "number",
                "operation": "gt"
              },
              "leftValue": "={{ $json.total_at_risk }}",
              "rightValue": 0
            }
          ]
        }
      },
      "typeVersion": 2.2
    },
    {
      "id": "f0630bc4-3158-4502-8d78-cdb6e891f709",
      "name": "Print \u2014 nothing to cut",
      "type": "n8n-nodes-base.code",
      "position": [
        35776,
        5712
      ],
      "parameters": {
        "jsCode": "// Nothing crossed the waste thresholds. Keep the shape the report expects.\nconsole.log('No delivery slices exceeded the waste thresholds. Nothing to cut.');\nreturn [{ json: { summary: 'No wasted spend found', ranked: [] } }];"
      },
      "typeVersion": 2
    },
    {
      "id": "2bec828b-29b1-44af-b943-69e80d8aa233",
      "name": "Set config: result events, floor",
      "type": "n8n-nodes-base.code",
      "position": [
        33248,
        5456
      ],
      "parameters": {
        "jsCode": "// Tune the finder here. RESULT_ACTIONS is every event that counts as a \"result\"\n// (leads, registrations, purchases) - the workflow sums whichever are present,\n// so it works for lead-gen and ecommerce without code changes.\nreturn [{ json: {\n  DATE_PRESET: 'last_30d',\n  SPEND_FLOOR: 300,        // ignore slices below this spend\n  CPR_MULTIPLE: 2,         // flag a slice if its cost-per-result is this many times the account benchmark\n  MIN_ATTRIBUTION: 0.5,    // skip a breakdown if it captures under this share of account results (attribution gap)\n  RESULT_ACTIONS: [\n    'offsite_complete_registration_add_meta_leads', 'lead', 'leadgen_grouped',\n    'onsite_conversion.lead_grouped', 'omni_complete_registration', 'complete_registration',\n    'omni_purchase', 'offsite_conversion.fb_pixel_purchase', 'purchase',\n  ],\n} }];"
      },
      "typeVersion": 2
    }
  ],
  "active": false,
  "settings": {
    "binaryMode": "separate",
    "executionOrder": "v1"
  },
  "versionId": "739f7aa0-b82d-4488-bbfd-853fbbaf403b",
  "connections": {
    "Run manually": {
      "main": [
        [
          {
            "node": "Set config: result events, floor",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Compute benchmark": {
      "main": [
        [
          {
            "node": "List delivery breakdowns",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Any budget at risk?": {
      "main": [
        [
          {
            "node": "Build AI prioritization prompt",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Print \u2014 nothing to cut",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Flatten & score slices": {
      "main": [
        [
          {
            "node": "Flag underperforming slices vs benchmark",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "List delivery breakdowns": {
      "main": [
        [
          {
            "node": "Fetch breakdown insights (Meta)",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Fetch account totals (Meta)": {
      "main": [
        [
          {
            "node": "Compute benchmark",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Build AI prioritization prompt": {
      "main": [
        [
          {
            "node": "Rank the kill list with Claude AI",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Fetch breakdown insights (Meta)": {
      "main": [
        [
          {
            "node": "Flatten & score slices",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Set config: result events, floor": {
      "main": [
        [
          {
            "node": "Fetch account totals (Meta)",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Rank the kill list with Claude AI": {
      "main": [
        [
          {
            "node": "Print prioritized wasted-spend report",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Flag underperforming slices vs benchmark": {
      "main": [
        [
          {
            "node": "Any budget at risk?",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}

About this workflow

This workflow pulls the last 30 days of Meta Ads insights across multiple breakdowns, calculates an account cost-per-result benchmark, flags high-waste delivery slices, and uses Anthropic Claude to rank a prioritized “kill list” with recommended actions. Runs when triggered…

Source: https://n8n.io/workflows/17012/ — original creator credit. Request a take-down →

More Social Media workflows → · Browse all categories →

Related workflows

Workflows that share integrations, category, or trigger type with this one. All free to copy and import.