Determines the combination of levels that make the optimal page
or "champion" of the test experiments. The Results report is available for multivariate
projects and tests.

## Predicted Optimal

A summarized look at the winning split page or multivariate level per factor.

**For Split Tests**

- Predicted Conversion Rate The percentage of test viewers
to activate a conversion event based on the views and conversions data collected
so far.
- Confidence Interval The amount of error in a predicted
value. Confidence interval shrinks as the number of test views grows.
- Statistically Significant Whether the result for any one
split page are statistically significant when compared to control. More test
views increases the likelihood of statistical significance.

**For Multivariate Tests**

- Factor The page element identified for testing.
- Significance The significance of the factor in the
results. High significance means that factor had a large impact on the results.
Only factors of High significance are included in the conversion rate
calculations.
- Content Level A list of content per factor with the best
performing content selected by default. Change the selection to evaluate how
different content combinations performed against control.
- Level Effect The percentage point (pp) deviation from the overall
mean conversion rate.
- Conversion Rate Effect The percentage point (pp) difference
between the effect the selected level will have on the conversion rate compared
to control. Calculated: (level influence of the selection + mean level
influence) - (control level influence + mean level influence)
- Lift Effect The effect a level will have on the lift
compared to control.

## Factor Results

A look at the performance of each level in the indicated factor.

- Level Influence with Mean Adds the level influence to the
mean influence to produce positive values for all content levels. The highest
influence generally indicates the "winning" content.
- Conversion Rate Effect The effect a level will have on
the conversion rate compared to control. Calculated: (level influence of the
selection + mean level influence) - (control level influence + mean level
influence)