Skip to main content
Confidence Cloud is the default for every new account, and it needs no setup. Confidence Cloud is a data warehouse that Confidence creates and operates for your account. It lets you define entities, send events, and run experiments without setting up warehouse infrastructure or credentials. The Console marks the warehouse as Managed by Confidence. Optionally, you can connect your own warehouse on BigQuery, Snowflake, Redshift, or Databricks. That keeps your data inside your infrastructure and lets you define tables with your own SQL. You can do it at any time, and the warehouse you connect becomes the primary warehouse of the account.

​
Confidence Cloud Compared to Your Own Warehouse

​
Demo Data

The demo data is a worked example of a checkout flow, so metrics and experiment results have something to show before your own events arrive: The dimension table carries a country dimension, the client comes with a credential named demo credential, and the Skip button rollout reaches 30% of users. Demo data assignment table belongs to the demo data, separate from the account’s own Confidence Cloud Assignment Table, and it reads only the mocked rows. The Console marks every one of these resources with a Demo data tag, beside the name in a list and beside the title on the resource’s own page, so you can tell them from resources of your own at a glance. The tag comes from the label demo_data, which the list methods of the API take as a filter, labels.demo_data:true. The rows behind the tables live in the confidence dataset of your warehouse project, in tables carrying a demo_ prefix, next to the tables your own data lands in. The experiment results come from mocked rows rather than from traffic, so they never change. Confidence can remove every labeled resource again when you no longer want them, from Exit demo mode in the Console.

​
How Data Reaches the Warehouse

Two built-in connections deliver all data:
  • Events that you send with the SDKs or the events API.
  • Flag assignments from flag resolves, when you use Confidence Flags.
You don’t manage either connection, and you don’t load tables of your own into the warehouse.

​
Metrics Without SQL

On Confidence Cloud, Confidence generates fact tables for you:
1

Define an entity

An entity represents what you measure, such as a user, an account, or a device.
2

Create an event that references the entity

Confidence creates a fact table for that event as soon as the event exists.
3

Create metrics on the fact table

Numeric event fields become measurements, the remaining fields become dimensions, and the event timestamp becomes the timestamp column.
An event that references no entity gets no fact table, because a fact table needs an entity to join assignments on.

​
Limitations

Confidence owns the warehouse, so the parts of the platform that write SQL against it are unavailable:
  • Fact, dimension, assignment, exposure, and entity relation tables that you define yourself. Confidence generates them from your events instead.
  • SQL previews and data exports.
  • Metric definitions kept in a repository and synced with metrics as code.
  • Materialized segments.
You can’t edit or delete the warehouse connection, and it stays the primary warehouse of the account.

​
Connect Your Own Warehouse Instead

You can connect your own warehouse at any time, on a plan that includes Warehouse Native Confidence. Connect it from Admin > Connections > Data warehouse, and Confidence then makes it the primary warehouse of the account. Confidence Cloud stays in the list as a read-only connection.