Introducing Confidence Cloud

Imagine you want to test a product idea, but your team does not have a data warehouse. Or perhaps your organization has one, but accessing it means waiting for help from another team.

Before you can learn from an experiment, you first have to provision infrastructure, configure credentials, and connect everything to your experimentation platform.

Today, we're introducing a faster way to get started: Confidence Cloud.

Confidence Cloud gives you a data warehouse created and operated by Confidence. You can define entities, send events, create metrics, and run experiments without setting up warehouse infrastructure or credentials of your own.

#From a new workspace to something worth exploring

A new experimentation tool can feel a little empty before your first event arrives. Confidence Cloud gives you something useful to explore from the start.

Your workspace includes clearly labeled demo resources that show how the pieces of Confidence work together: flags, metrics, A/B tests, and rollouts. The demo results use sample data, so you can explore a realistic product workflow without generating traffic first.

When you're ready to work with your own product, connect your application and begin sending events.

#From events to experiment results

Confidence Cloud is built around the events you send from your application.

Start by defining who or what you measure: such as a user, account, or device. Then define the events that describe product behavior, such as product-viewed, added-to-cart, or checkout-completed.

Confidence automatically creates the fact tables needed to build metrics from supported event data. Those metrics can then measure the effect of experiments, rollouts, and other product changes.

If you use Confidence Flags, flag assignments are delivered automatically too. This connects what users experienced with what they did next, without requiring you to build a separate assignment-data pipeline.

#Start building

Choose the setup path that works best for you:

• Use the CLI wizard: npx @spotify-confidence/quickstart
• Let your coding agent use the Confidence agent skills
• Follow the product tours in the Confidence Console

These guides help you define the entities you measure, the events that describe product behavior, and the metrics that tell you whether a change worked. They also guide you through integrating events and feature flags into your application.

As data begins to arrive, you can use the Confidence Agent to explore product behavior and investigate experiment results. You can also connect quantitative findings with relevant Session Recordings to understand not only what changed, but why.

#Confidence Cloud or warehouse-native Confidence?

Confidence now supports two ways to run the data layer behind your experiments.

Confidence Cloud is the streamlined path. It is designed for teams that want to start experimenting without first assembling warehouse infrastructure or coordinating access with a data platform team.

Warehouse-native Confidence gives you more flexibility and control. You can use existing business data, define tables with custom SQL, and keep your own warehouse and data models as the source of truth. It remains the best fit for organizations with established data infrastructure or more advanced data requirements.

Same Confidence platform, just two ways to get started.

#Questions you might have

Is warehouse-native Confidence going away?
No. Confidence Cloud just adds a faster starting path; it does not replace warehouse-native Confidence. Confidence continues to support customer-managed BigQuery, Snowflake, Redshift, and Databricks warehouses. Warehouse-native Confidence remains the most flexible option for organizations that need custom SQL, existing business datasets, or greater control over their data infrastructure.

What does Confidence Cloud run on?
Confidence Cloud uses BigQuery. Confidence manages the warehouse infrastructure, storage, and compute, so you do not need to create a Google Cloud project or configure warehouse credentials.

How is Confidence Cloud different from warehouse-native Confidence?
Confidence Cloud is streamlined around events sent to Confidence and assignments generated by Confidence Flags. Confidence creates the tables needed to build metrics from those events. Warehouse-native Confidence lets you connect existing data, define tables with custom SQL, and manage the warehouse configuration yourself. It requires more setup, but gives your organization more control over its data and infrastructure.

Can I use my own warehouse?
Yes. Confidence supports customer-managed BigQuery, Snowflake, Redshift, and Databricks warehouses. Upgrade to the growth plan or contact us if warehouse-native Confidence is the better fit for your organization.

#Spend less time setting up and more time learning

Confidence Cloud removes warehouse setup from the path to your first experiment. You can explore a realistic demo, connect your application, turn product events into metrics, and start learning from experiments, without first provisioning warehouse infrastructure.

Less time assembling experimentation infrastructure. More time building, measuring, and learning.

Start building with Confidence.

#Further reading

• Introducing Confidence Agent: the AI collaborator that works with your flags, experiments, and metrics
• Introducing Recordings: replaying real user sessions to understand why a metric moved
• A/B Tests and Rollouts: the distinction between testing ideas and safely releasing changes
• A/B Testing Bandwidth: why experiment velocity is the binding constraint on innovation