From Marketing Data in BigQuery to an Interactive Dashboard — In a Single Conversation with Claude

You are running campaigns across Meta Ads, Google Ads, TikTok, LinkedIn, and more. The data is fragmented across platform dashboards that each tell a different story. Getting a unified view — ROAS by channel, spend vs. revenue trend, cross-platform CAC — takes hours of exporting, cleaning, and copy-pasting into spreadsheets.

What if you could consolidate all of that marketing data into a single BigQuery table automatically, and then open Claude, ask a question in plain English, and say “now build me an interactive dashboard from this” — and it just did?

That is exactly what this guide covers. A complete pipeline: Porter Metrics pushes your marketing data into BigQuery on autopilot, and Claude queries that data and generates the dashboard — all in one conversation.

The Architecture: Five Nodes, One Conversation

Before touching any configuration, it helps to understand what we are building. The full pipeline has five components that work together as a seamless chain:

That is the entire chain. Let us build it — starting with the most important step: getting your marketing data into BigQuery.

Step Zero: Load Your Marketing Data into BigQuery with Porter Metrics

Before Claude can query anything, your marketing data needs to be in BigQuery — structured, normalized, and automatically refreshed. This is the step most tutorials skip, and it is the reason most AI-generated dashboards end up being demos rather than production tools.

Manually exporting CSVs from Meta Ads, Google Ads, and TikTok and uploading them to BigQuery is not sustainable. You need a pipeline. Porter Metrics is that pipeline — a no-code connector that pulls data from 20+ marketing platforms and writes it directly into your BigQuery tables on a schedule.

Step-by-Step: Send Marketing Data to BigQuery with Porter Metrics

Prerequisites for the Claude MCP Connection

With your marketing data now in BigQuery, here is what you need to connect Claude to it.

Step 1: Create a Read-Only Service Account for Claude

We need a dedicated Google Cloud identity for Claude — a service account with read-only access to your marketing data tables. This keeps Claude’s access scoped and auditable.

Create the Service Account

Assign the Minimum Required Roles

Download the JSON Key File

# Move to a secure directory
mkdir -p ~/.credentials
mv ~/Downloads/your-key-file.json ~/.credentials/claude-bigquery.json

# Restrict file permissions (macOS / Linux)
chmod 600 ~/.credentials/claude-bigquery.json

Step 2: Configure the BigQuery MCP Server

The Model Context Protocol (MCP) is an open standard that lets Claude talk to external systems through a defined set of tools. The BigQuery MCP server exposes three core tools to Claude: list_tables, describe_table, and execute_query.

Edit the Claude Desktop Configuration File

# macOS
open ~/Library/Application\ Support/Claude/claude_desktop_config.json

# Windows
notepad %APPDATA%\Claude\claude_desktop_config.json

Add the following block inside mcpServers:

{
  "mcpServers": {
    "bigquery": {
      "command": "uvx",
      "args": ["mcp-server-bigquery"],
      "env": {
        "BIGQUERY_PROJECT": "your-gcp-project-id",
        "BIGQUERY_LOCATION": "US",
        "BIGQUERY_KEY_FILE": "/Users/yourname/.credentials/claude-bigquery.json"
      }
    }
  }
}

Restart Claude Desktop

Close Claude Desktop completely and reopen it. On the new conversation screen, you should see a small tools icon in the input area — this confirms that Claude has detected and loaded the MCP server.

Step 3: Verify the Connection with 3 Marketing Test Queries

With the MCP server running, open a new chat in Claude Desktop and run these three verification queries. They confirm the pipeline is working end-to-end with your actual marketing data.

Step 4: The Master Prompt for Marketing Dashboard Generation

This is the step that changes everything for marketers. Most guides stop at “Claude can answer questions about your data.” What this pipeline does is go further: Claude queries your BigQuery marketing data live and then generates a complete interactive dashboard — all in the same conversation.

Use this prompt structure as your starting template:

You have access to my BigQuery marketing data through the MCP connection.

STEP 1 — Data retrieval:
Query the [project.dataset.table] table and pull [metric_1], [metric_2], [metric_3]
for the last [time period], grouped by [platform / campaign / date].
Also calculate [derived_metric] if relevant (e.g., ROAS = revenue / spend).

STEP 2 — Dashboard generation:
Using the data you just retrieved, build a complete interactive dashboard as a React Artifact with:
- A header with the dashboard title, date range, and last-updated date
- KPI cards at the top: [Total Spend], [Total Revenue], [Overall ROAS], [Total Conversions]
- A line chart showing [spend / revenue] trend over time, broken down by [platform]
- A horizontal bar chart ranking [platforms / campaigns] by [ROAS / revenue]
- A comparison table with all platforms: Spend, Revenue, ROAS, CAC, Conversions
- A filter to switch between platforms or date ranges
- Clean, professional design — [light / dark] color scheme

AUDIENCE: [CMO / agency client / performance marketing team]
PURPOSE: [weekly review / client report / budget allocation meeting]

Step 5: Publish and Share Your Marketing Dashboard

Once the dashboard renders in the Artifact panel, publishing it takes a single click. In the top-right corner of the Artifact panel, click the Publish button. Claude generates a public URL that opens your fully interactive marketing dashboard — no login required for viewers.

Share the link in:

• Slack — Drop it in the channel for async team review
• Client reports — A live, interactive dashboard is far more impressive than a static PDF or screenshot
• Agency presentations — Open it full-screen during the monthly performance call
• Notion or Confluence — Embed the URL as an iframe in your campaign documentation

Marketing Use Cases: Real Workflows Teams Are Running Today

The setup is the same for every team. The use cases are where the real value emerges. Here is how marketing teams and agencies are using this BigQuery + Claude pipeline in practice.

BigQuery + Claude vs Traditional Marketing BI Tools

The question marketers ask most often: why use this pipeline when Data Studio is free and Looker (formerly Looker Enterprise) already exists? Here is an honest comparison focused on marketing team realities.

The honest take: Data Studio is fine for simple, single-platform reports. Looker is powerful but requires a data engineering team and a five-figure annual contract to justify. BigQuery + Claude via Porter Metrics is the gap in between — powerful enough for serious marketing analysis, fast enough for weekly reporting, and cheap enough for teams of any size.

The “Live Dashboard” vs the “Dead Dashboard”

Traditional marketing dashboards in Data Studio are dead by design — they answer the questions the builder anticipated when they built them, and nothing more. Want to slice campaign performance by a new dimension? That is a ticket to the data team or 30 minutes rebuilding a chart.

The BigQuery + Claude pipeline produces something qualitatively different: a live conversation around your marketing data. After Claude generates your initial dashboard, you can immediately continue:

• “Break the ROAS chart down by campaign objective instead of platform.”
• “Add a column showing spend as a percentage of total budget.”
• “Which campaigns should I pause based on this data? Give me a recommendation.”
• “What is driving the drop in conversion rate in week 8? What could explain it?”

Each instruction updates the Artifact in real time and Claude can also provide written analysis — not just charts. That combination of visual dashboard plus AI narrative is something no static BI tool can replicate.

Security: What You Must Get Right

Connecting an AI tool to your marketing data warehouse requires basic security discipline. Here are the non-negotiable practices for this setup:

• Keep the service account strictly read-only: The Claude service account should only have the three IAM roles listed in Step 1 — and access only to the datasets created by Porter Metrics. Never grant project-level admin permissions.
• Never commit the JSON key file to version control: Add .credentials/ to your .gitignore immediately. If you use a shared team environment, consider GCP Workload Identity Federation instead of key files.
• Restrict to SELECT only: If your MCP server implementation supports it, explicitly block INSERT, UPDATE, DELETE, DROP, and ALTER. Claude should read your marketing data — never modify it.
• Audit Claude’s query activity: Every query Claude runs is logged in BigQuery’s INFORMATION_SCHEMA. Review JOBS_BY_PROJECT periodically to audit what has been queried and confirm Claude is only accessing the marketing tables you intend.

-- Audit recent queries run by Claude's service account
SELECT
  job_id,
  creation_time,
  user_email,
  query,
  total_bytes_processed,
  state
FROM `region-us`.INFORMATION_SCHEMA.JOBS_BY_PROJECT
WHERE user_email = 'claude-bigquery-readonly@your-project.iam.gserviceaccount.com'
  AND creation_time > TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 7 DAY)
ORDER BY creation_time DESC
LIMIT 50;

Next Steps: Where to Take This Further

Conclusion: Marketing Intelligence That Moves at the Speed of Your Questions

For years, the bottleneck in marketing reporting has not been the data. It has been the translation layer — the data team ticket, the dashboard developer, the LookML model, the Data Studio chart rebuild. By the time the report lands in front of you, the campaign decision has already been made.

The Porter Metrics + BigQuery + Claude pipeline collapses that translation layer. Your marketing data flows automatically from 20+ platforms into BigQuery. Claude reads it, analyzes it in plain English, and generates an interactive dashboard you can publish and share in minutes. The marketer with the question is now the marketer who gets the answer — on demand, without a data team in the loop.

Start with one channel. Connect Meta Ads or Google Ads to BigQuery via Porter Metrics, run the three verification queries in Step 3, and use the master prompt in Step 4 to generate your first dashboard. The first time it works, you will immediately understand why this approach is replacing traditional BI for marketing teams.