🚰 Feedback

You can now export experiment results as Tabular CSVs—built specifically for clean imports into tools like Airtable, Notion, Coda, and Monday. CSV Aggregated Tabular: One row per variation/goal with totals for the selected date range CSV Daily Tabular: Daily time-series data with a flat, predictable schema Both formats use header-first tables, consistent column names, and machine-friendly data types, so you can import them directly—no cleanup required. Perfect for recurring reports, dashboards, and automated workflows. Learn more: https://support.convert.com/hc/en-us/articles/export-download-in-convert-understanding-the-different-export-types-

You can now connect your Convert optimization data directly to Claude Desktop, Cursor, or any MCP-compatible AI assistant—no middleware, no complicated setup. With the Convert MCP Server, you get: > Zero-config setup (run it via npx) > Full REST API v2 coverage (organized into clean namespaces) > Built-in Convert knowledge (5,279+ indexed Help Center chunks) > Configurable access levels: reporting (default), readOnly, or all (read/write) Once connected, you can ask in plain English: “Show all active experiments and performance” “Summarize top winners from the last 30 days” “Which audiences are performing best?” Start with reporting access and upgrade permissions only if you need deeper configuration or write actions. Learn more here: https://support.convert.com/hc/en-us/articles/convert-mcp-server

All new experiments now use t-tests for Frequentist statistics (experiments created before Oct 13, 2023 continue using z-tests). You can configure key settings based on how critical your test is, including confidence level (95% or 99%), test type (two-tailed by default, one-tailed optional), and multiple comparison correction (Sidak, Bonferroni, or none). To make setup easier, Sensible Defaults let you quickly apply recommended settings for standard or mission-critical experiments. Learn more: https://support.convert.com/hc/en-us/articles/10157554451085-statistical-methods-used

We’re adding Risk and Expected Uplift to Bayesian reports to help you make faster, more confident decisions. See the realistic impact a variant could deliver and the downside to watch out for; alongside Chance to Win. Great for high-traffic tests, revenue-critical flows, early stopping, and comparing close winners. Metrics are optional and can be toggled on or off. Learn more: https://support.convert.com/hc/en-us/articles/understanding-risk-and-expected-uplift-in-bayesian-reports