How to Use Looker with Google Cloud Databases

Connecting Looker to your Google Cloud databases like BigQuery or Cloud SQL is a surefire way to build a powerful and scalable analytics engine for your business. Instead of having your data stuck in a complex warehouse, you give your entire team a user-friendly way to ask questions, build dashboards, and uncover valuable insights. This article will guide you through connecting Looker to BigQuery and Cloud SQL and share some best practices to get you started on the right foot.

Why Pair Looker with Google Cloud?

Modern businesses run on data, and the Google Cloud Platform (GCP) offers some of the most powerful and scalable database solutions available. Looker acts as the user-friendly interface or "semantic layer" on top of that raw power. Think of your Google Cloud database as a high-performance car engine and Looker as the dashboard, steering wheel, and navigation system - it makes all that power accessible and useful to anyone, not just the mechanics (or data engineers).

Here’s what makes this combination so effective:

• Scalability: Google BigQuery can handle petabytes of data effortlessly. As your business grows, your data stack won't slow you down. You can analyze years of marketing data, millions of customer records, or massive event logs without a hiccup.
• Unified View of Your Business: By centralizing data from various sources (your website, CRM, ad platforms) into BigQuery, you can use Looker to create a single source of truth. No more comparing conflicting numbers from different department spreadsheets.
• Real-time Insights: Connect Looker to a live database like Google Cloud SQL and you get dashboards that reflect what's happening in your business right now. Monitor sales activity, application performance, or operational metrics as they happen.
• Security and Governance: Both Looker and Google Cloud provide robust tools for managing data access. You can control exactly who sees what data, right down to the specific rows and columns they're allowed to view.

The Core Components Explained

Before diving into the setup, let's quickly clarify the roles of the key players: Looker, BigQuery, and Cloud SQL. Understanding what each does will help you decide which one to use.

Looker

Looker is much more than just a dashboard tool. Its core strength lies in its modeling layer, LookML. With LookML, your data team can define all your business metrics, calculations, and data relationships in one place. For example, they can create a single, official definition for "Monthly Active User" or "Customer Lifetime Value." This means your marketing, sales, and product teams can all build their own reports using reliable terminology without needing to write a single line of SQL. They are exploring governed, trusted data, ensuring everyone is working from the same playbook.

Google BigQuery

BigQuery is a fully-managed, serverless data warehouse. That’s a technical way of saying it's a massive, powerful database built specifically for running fast analytical queries on huge datasets. You send data to it - from Google Analytics, Facebook Ads, Shopify, Salesforce - and it stores it in a way that makes analysis incredibly fast. Use BigQuery when: You need to analyze large volumes of historical data from multiple sources to understand trends, run campaign performance analysis, or build complex business intelligence reports.

Google Cloud SQL

Google Cloud SQL is a fully-managed service for running relational databases like PostgreSQL, MySQL, and SQL Server. Relational databases are what power most applications - they’re excellent at handling transactions and storing structured data in a highly organized way, like a perfect digital filing cabinet. Use Cloud SQL when: You want analytics on live operational data. For instance, running reports directly off your e-commerce store's backend database, your SaaS application database, or any system with constantly changing transactional information.

Connecting Looker to Google BigQuery: A Step-by-Step Guide

Connecting Looker to BigQuery involves creating a secure access credential in Google Cloud and then plugging it into Looker. Here’s a breakdown of the process.

Step 1: Get Your Google Cloud Service Account Credentials

First, you need to create a dedicated “service account” in your Google Cloud project. This is essentially a robot user that Looker will use to access your data. Using a dedicated account is a crucial security practice so that you can control permissions finely.

1. Navigate to "IAM & Admin" in your Google Cloud Console and select "Service Accounts."
2. Click "Create Service Account." Give it a descriptive name like looker-bigquery-connector.
3. Grant it the necessary IAM roles. For Looker to work properly, you’ll typically need two roles:
4. Click "Done" to create the account. Now navigate back to the service account, click on the "Keys" tab, and select "Add Key" > "Create new key."
5. Choose "JSON" as the key type and click "Create." A JSON file will automatically download. Treat this file like a master password. Store it securely and never expose it publicly.

Step 2: Add the BigQuery Connection in Looker

With your JSON credential file in hand, head over to your Looker instance.

1. Go to the Admin section and click on Database > Connections in the left sidebar.
2. Click the "Add Connection" button.
3. Configure the connection settings:
4. (Optional but Recommended) Enable Persistent Derived Tables (PDTs): This is a powerful feature where Looker can pre-calculate complex tables, making your dashboards much faster. To enable it, check the "Persistent Derived Tables" box and specify a "Temp Dataset." This is a scratch dataset in BigQuery where Looker can write temporary tables. You must grant the service account BigQuery Data Editor permissions on this dataset.

Step 3: Test and Finalize

Before saving, click the "Test" button at the bottom of the page. Looker will attempt to connect to BigQuery using your credentials. If everything is configured correctly, you'll see a success message. If not, double-check your service account permissions and project ID for typos. Once it passes, click "Add Connection" and you're ready to start building!

Connecting Looker to Google Cloud SQL

Connecting to Cloud SQL is more like connecting to a traditional database. The main difference is ensuring network access is correctly configured.

Step 1: Configure Your Cloud SQL Instance for Access

For security, Cloud SQL instances are not publicly accessible by default. You need to explicitly grant Looker permission to connect.

1. First, you’ll need to create a dedicated read-only database user for Looker within your Cloud SQL instance. This follows the principle of least privilege - Looker only needs to read your data, not change it. Note the username and password.
2. Next, you need to authorize Looker’s IP addresses. In the Google Cloud Console, navigate to your Cloud SQL instance.
3. Go to the "Connections" tab, and under "Authorized networks," click "Add network."
4. You'll need to enter Looker's egress IP addresses here. You can find the current list in Looker’s official documentation by searching for "Looker IP addresses." Add each one to the authorized list.

Step 2: Add the Connection in Looker

This part will feel familiar.

1. Go to Admin > Database > Connections and click "Add Connection."
2. Configure the settings:
3. Ensure the "Verify SSL Cert" option is checked for a secure connection.

Click "Test," and on success, click "Add Connection." You're now connected to your live operational database.

Best Practices for a Smooth Experience

Just connecting your data is only the beginning. To get the most out of your Looker and Google Cloud setup, keep these tips in mind:

Start with Your LookML Model

The real magic of Looker is LookML. Before you rush to build a dozen dashboards, take the time to define your primary metrics and business logic. Create dimensions and measures for concepts like "customer," "revenue," and "session." This upfront investment ensures consistency and frees your business users from worrying about the underlying complexity.

Optimize Cost and Performance in BigQuery

BigQuery charges for the amount of data processed by your queries. To keep costs down and dashboards fast, use BigQuery features like partitioning and clustering on your large tables. In Looker, leverage features like aggregate awareness and PDTs to query smaller, pre-aggregated tables instead of scanning massive raw tables every time a dashboard loads.

Manage Permissions Deliberately

Use Looker groups and roles to manage who can see and do what. You can restrict access to certain models (e.g., the finance team only sees financial data), explores, or even specific columns. This keeps sensitive data secure and prevents users from being overwhelmed by irrelevant information.

Iterate and Expand

Don't try to boil the ocean on day one. Start by connecting a single high-value dataset and answering a few critical business questions for one team. Build a successful proof of concept and get them excited. This success will generate pull from other departments, allowing you to iterate and expand your Looker usage organically.

Final Thoughts

Connecting Looker to Google Cloud databases is a game-changer, turning your mountains of raw data into a clear, explorable landscape for your entire organization. By pairing Looker's powerful data modeling with the scalability of BigQuery and the real-time nature of Cloud SQL, you build a foundation for making smarter, faster, data-informed decisions.

While the process is powerful, setting up and modeling data in Looker still takes time and technical expertise. We know from experience that many teams simply don't have the hours to become BI experts, they're busy running marketing campaigns and closing deals. That’s why we created Graphed. Our platform automates the entire process: connect your sources like Google Analytics, Shopify, and Salesforce with one click, and then use plain English to build the real-time dashboards you need in seconds. Instead of a multi-day setup project, you can get from data to insights in the time it takes to write an email.