Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.kameleoon.com/llms.txt

Use this file to discover all available pages before exploring further.

Controlled-experiment Using Pre-Experiment Data (CUPED) uses data from existing visitors to accelerate reliable experimentation and personalization results.

​
Key benefits

Use CUPED to make data-driven decisions faster and with more confidence. CUPED provides better estimates for variation effects and reduces the risk of false-positive results for tests with limited data.

​
When to use CUPED

CUPED is most effective when:
  1. You have collected goal data for at least two weeks prior to the experiment launch.
  2. A correlation exists between goal conversions before and during the experiment.
Consider your experiment goals before enabling CUPED. Activate CUPED for the main metrics of your site, such as transactions.

​
Enable CUPED

To enable CUPED:
  1. Navigate to the Data tab in the right-hand sidebar.
  1. Under Advanced settings, click the toggle next to CUPED.
  1. Click the dropdown menu, then select the goals for which you want to enable CUPED.
  2. Click Apply.
Data loading might take longer than the classic algorithm. Limit CUPED to your key goals to maintain performance.
The page data changes based on the selected goals. If you select your primary goal, the main results block displays CUPED-adjusted data and is labeled CUPED enabled.
Similarly, the results tables for each goal update based on the enabled CUPED settings.
The results columns display three primary differences:
  • Conversion rate becomes Conversion rate adjusted
  • Improvement rate becomes Improvement rate adjusted
  • Reliability becomes Reliability adjusted
Apply filter-based breakdowns to results data only if you disable CUPED. When you click Breakdown on a table, you must confirm you want to disable CUPED. You can reactivate CUPED after removing the breakdown filter.
CUPED effectiveness increases over time. As you use Kameleoon, the algorithm leverages more experiment data to improve accuracy.

​
Manage CUPED

To modify your CUPED configuration:
  1. Open the Actions menu.
  2. Select Manage CUPED.
Select or deselect goals in the pop-up window, then click Validate.

​
Disable CUPED

To revert to the classic statistical method:
  1. Open the Actions menu.
  2. Select Manage CUPED.
Deselect the goals in the pop-up window and click Validate. CUPED no longer applies to your results.
Once you enable CUPED, rely on its results consistently. Switching between original and CUPED-adjusted results to find more favorable outcomes increases the risk of overestimating lift and encountering false positives.
For more information on CUPED, read the Kameleoon statistical paper.