Everything we ship, in order. Subscribe to the RSS feed or just keep this tab open — most weeks there is something new.
Every forecast panel now renders an 80% and 95% confidence interval instead of a single projected line, so you can see how uncertainty widens further out.
Fixed a rounding error that could show retention percentages up to 0.4 points higher than actual for cohorts with fewer than 200 users.
Reverse ETL syncs to Snowflake and BigQuery are now up to 3.4x faster thanks to incremental change-data-capture instead of full-table scans on every run.
Set different attribution lookback windows per product category — useful for teams where a 3-day window fits impulse categories and a 30-day window fits considered purchases.
Anomaly alerts that fired on the same underlying incident within a 15-minute window are now grouped into a single notification instead of paging your on-call channel repeatedly.
Define a metric once — like "active user" or "qualified lead" — and every dashboard, alert and scheduled report in the workspace now inherits the same definition automatically.
A rebuilt query planner cut median dashboard load time from 1.8s to 0.7s across workspaces with more than 50 saved charts.
Enterprise workspaces can now manage seats through SCIM, syncing user creation, role assignment and deprovisioning directly from Okta or Azure AD.
Scheduled reports now consistently render in the workspace timezone instead of falling back to UTC for recipients outside the account owner’s locale.
Every workspace can subscribe its own Slack channel to release notes — no separate mailing list to manage.