2025 Optimizely Analytics release notes
8 min
Follow this article to receive email notifications about Optimizely Analytics and Warehouse-Native Experimentation Analytics updates for 2025.
December 2025
- Metric presets – Apply consistent formatting to metrics across all analyses. Set currency symbols, decimal places, and number formatting once at the metric level, and ensure consistency across explorations, dashboards, and experiment results.
- Attribution block – Understand which user actions and touchpoints drive conversions with attribution blocks. Analyze the customer journey and attribute value to each interaction across channels and campaigns.
- Monitoring service – Get visibility into the health and status of your application's background processes with the task monitoring service, ensuring your data is always accurate and up-to-date.
- CUPED in Redshift – Reduce experiment noise and detect smaller effects faster with CUPED (Controlled-experiment Using Pre-Experiment Data), now available for Redshift customers.
- Saved ratio and conversion rate metrics – Calculate and analyze ratio metrics like conversion rates with built-in metric templates, eliminating the need for custom formulas.
November 2025
Subprocessor update on November 12, 2025
Optimizely Analytics will migrate its managed infrastructure from Amazon Web Services (AWS) to Google Cloud Platform (GCP) no sooner than December 12, 2025, with completion anticipated in Q1 2026. This change will enhance performance, scalability, and reliability while maintaining existing data protection and security commitments. No action is required from customers, except those whitelisting IPs — we will send follow-up instructions directly. There will be no change in how Optimizely transfers, secures or handles customer data. The hosting region will remain in the same national jurisdiction.
October 2025
- AI Exploration Generator – Generate explorations instantly by typing natural-language questions. This feature is only available to customers on Opti ID.
- Favorites – Mark key entities (explorations, dashboards, etc.) as favorites to keep them easily accessible, so you can quickly return to the insights you use most, without digging through folders or searching.
- 'Compare to' tile in Dashboards – Perform period-over-period comparisons across all exploration tiles within a dashboard using the Compare to tile in Optimizely Analytics dashboards.
- Eager loading in Dashboards – Enable proactive loading of all dashboard tiles – visible or not.
- Advanced configuration in Connections – Add custom key-value pairs in Advanced Configuration for database connections beyond the main configuration parameters, for enhanced flexibility.
- Table totals and subtotals – Switch easily between the big picture and the details. Expand or collapse tables to see totals, subtotals, and detailed breakdowns.
Warehouse-Native Experimentation Analytics
- Guardrail alerts – Send notifications when a key metric declines beyond a set threshold during an experiment. Often tied to guardrail metrics like site speed or signups, these alerts help prevent negative outcomes and enable fast, proactive decision-making.
July 2025
- Categories – Organize your workspace by assigning entities such as dashboards and events to a category. This makes it easier to find, manage, and govern your workspace at scale. Assign categories by team or product, or designate them as “verified” to reduce clutter, establish ownership, and enable faster navigation.
- Analyzing experiments using categories – Analyze multiple metrics that share the same category by selecting the category itself. No need to manually select the same set of metrics for every experiment.
- Dashboard filter & parameter persistence – Share dashboards with filters and parameters pre-applied, making collaboration across teams seamless. Copy and send the URL, and when a user opens the URL, they see the same filtered view. This ensures AEs, CSMs, and other stakeholders have access to consistent information without additional steps.
- Save a metric or derived column from block shelf – Open and save a metric or derived column that you create within an exploration. This removes the need to recreate metrics or derived columns that you have already created in a block shelf.
For customers on Optimizely ID (Product switcher)
- AI chat – Ask Opal anything—from how to build an exploration to Analytics best practices and advanced Optimizely features. Opal has access to the full Optimizely documentation, so you can get answers to technical Optimizely Analytics questions without ever leaving the product.
- AI exploration summary – Summarize any exploration and surface key takeaways directly within Optimizely Analytics. Save time interpreting results and share insights more effectively with your team.
June 16, 2025
- Selector performance improvements – Load selectors faster, especially when working with a large number of entities such as columns, datasets, or metrics.
- Selector search improvements – Search efficiently, even with typos or partial names, to find what you need.
- Parameterized rolling window – Set rolling window lengths in moving aggregate formulas for greater flexibility.
- Percentile aggregate function – Calculate the value below which a chosen percentage of your data falls, helping to identify distribution patterns and understand typical versus extreme user behavior.
- Improved unsaved changes messaging – Keep track of unsaved edits with clear, actionable messages.
Warehouse-Native Experimentation Analytics
- Experiment on anything – Use Optimizely’s Stats Engine to analyze experiments from any source, as long as variation assignment data is in your warehouse.
- Outlier management – Boost result reliability by capping extreme values at configurable thresholds to prevent a few edge cases from skewing experiment outcomes.
- Sample ratio mismatch check – Detect imbalances in experiment variation traffic, helping troubleshoot implementation or instrumentation issues early.
May 8, 2025
- Visibility controls – Show or hide parts of your visualization without rerunning your query or changing your selection – especially useful for experiments with many metrics, variations, or crowded explorations.
- Rounding capabilities – Edit your visualization and round large values into millions, billions, or trillions to keep data readable. For example, render 1,500,000 as 1.5 M.
- Quarterly time grains – Use the Last tab of the date selector to group data by calendar or custom fiscal quarter and years.
- Quarter to date, month to date, week to date time ranges – Use QTD, MTD, or WTD presets from the Since tab of the date selector to apply common look-back windows.
- Key-pair authentication support for Snowflake – Support for key-pair authentication is now available alongside username-password login.
- % of Baseline – Compare how each variation performs against the baseline in the Experiment section of Event Segmentation, Universal, and Funnel templates.
- Conversion windows – Set a conversion window to control how long user behavior is counted after they’re assigned to a variation – for example, exclude actions that happen more than a day later to focus on immediate experiment impact.
- Ratio metrics – Create ratio metrics by dividing one event count or existing metric by another.
March 7, 2025
Warehouse-Native Experimentation Analytics is now generally available. The integration brings the elements of warehouse-native Optimizely Analytics to Feature Experimentation and Web Experimentation. Teams can analyze experiment performance, identify winning variations, and conduct deeper analyses on experiments that ensure data security and privacy and avoid data duplication or movement.
- Enhance experimentation results by integrating Optimizely Experimentation data with additional insights from the data warehouse. See scorecardscorecard for more information.
- Specify key user interactions to assess engagement and evaluate impact using custom eventscustom events.
- Create specific experiment-focused metrics (such as conversionconversion, numeric aggregationnumeric aggregation, ratioratio, and more).
- Segment users by common behaviors into cohortscohorts for precise analysis and targeted insights.
- Create custom metricsmetrics and derived columnsderived columns to transform data to gain deeper insights.
- Use the Stats EngineStats Engine to ensure reliable results and advanced analysis capabilities.
- Use CUPEDCUPED to reduce the impact of random variation and surface insights quicker.
- Switch effortlessly between configuring experiments and conducting deep experimentation analysis from both Feature Experimentation and Web Experimentation.
- Filter results by user segments, analyze trends over time, and track variation performance through designated funnels via Experimentation Analytics Experimentation Analytics > Explore.
- Uses the Opti ID Admin CenterOpti ID Admin Center for user management, giving you a single login point to switch among your Optimizely products. See the Opti ID documentationOpti ID documentation to learn more about how to use it.
- Updated the Analytics UI to match Optimizely styling.
Learn more about Warehouse-Native Experimentation Analytics.Learn more about Warehouse-Native Experimentation Analytics.
January 22, 2025
- Refreshed chart settings UI – Experience a smoother, more consistent chart editing process with the simplified chart settings interface with two streamlined tabs – Data and Style.
- Slowly changing dimension(SCD) tables – Use type 2 Slowly Changing Dimension (SCD2) tables to manage historical data changes more effectively with minimal configuration. The following are the key features:
- SCD2 datasets are modeled with three required fields:
- The actor dataset (commonly Users).
- The start time column, which must be a timestamp.
- The end time column, which must be a timestamp.
- Automatic integration – SCD2 datasets require only a many-to-one relationship with the actor dataset. Any event stream that joins with the actor dataset (for example, Users) will automatically join with the SCD2 table when configured.
- Simplified configuration – You do not need to create direct relationships between event and SCD2 datasets; analytics handles it seamlessly.