Build on Past Analyses with FactDecode
Published on October 27, 2026
Find Past Analyses More Easily
FactDecode lets you compare multiple Runs with different factor combinations and analysis settings. But as you execute more analyses, finding a particular Run worth revisiting can become time-consuming.
With this update, you can now bookmark completed Runs. Select a Run from the Results history to add or remove a bookmark. Bookmarked Runs display a dedicated icon, making them easy to distinguish from other analyses.
Run History now shows your 20 most recent Runs alongside all bookmarked Runs. Even as you create new analyses, bookmarked Runs remain easy to find in the default view. You can also continue to browse the full Run history whenever needed.
Bring a Past Run's Design into Your Next Analysis
When comparing historical results, you might want to revisit a previous factor combination or use it as the starting point for a new experiment.
FactDecode now lets you reuse the candidate factor configuration from a saved Run in your next analysis within the same Project. When available, select "Use this run's design" in Results to review the saved configuration and apply it to your current Design.
The restored configuration includes user-defined factor selections, Raw Data selections, and the maximum number of factors to select for the final model. Before applying it, a confirmation dialog shows the differences from your current Design. Changes are made only after you confirm, helping prevent unintended overwrites.
Once the Design has been restored, you can prepare another analysis using that configuration or modify the factors to explore a different combination.
Reanalyze with the Latest Available Data
Reusing a saved Run's Design is not the same as replaying the original analysis. Only the candidate configuration is carried forward—not the historical data, trained model, or prediction results.
Restoring a Design does not automatically start a new Run. Instead, you follow the normal analysis workflow, review the currently available data, and execute a new analysis. Even with the same factor configuration, results may differ because of updated data or changes in the analysis methodology.
In some cases, a saved Design may no longer be restorable—for example, if an older Run lacks the required configuration details or a factor's calculation definition has changed. FactDecode explains why restoration is unavailable rather than silently dropping factors and applying an incomplete configuration.
Clearer Results and More Helpful Explanations
We've also refined the presentation and explanations throughout Results. By reducing implementation-specific details and repetitive technical descriptions, FactDecode places greater emphasis on the information that matters when reviewing an analysis, including predictions, selected factors, and out-of-sample (OOS) validation metrics.
Help content has been simplified to explain what each metric represents and how to interpret it. Important distinctions, such as the difference between in-sample reference performance and historical OOS validation, remain available to support informed interpretation.
Your Saved Analysis Results Remain Unchanged
Bookmarking Runs or reusing their Designs does not recalculate or modify previously saved predictions and validation results. Your existing analyses remain available in Results.
FactDecode is designed for more than running an analysis once and moving on. Our goal is to help you compare results, preserve useful findings, and build on them in future analyses. With these improvements, it's easier to connect past research with your next experiment.