What is a holdout control group?
A holdout control is a share of your visitors who keep seeing the original page while new versions are tested. Because they arrive at the same time from the same sources, the difference between them and everyone else is the real effect of the change, not of seasonality, ads or luck.
Why it matters
Before-and-after comparisons are easy to fool: a sale, a new ad or a holiday can move conversion more than any page change. A control shown to the same traffic at the same time cancels those out. Without one, no tool can prove it raised sales.
How big should it be?
Big enough to measure, small enough to cost little. In a two-version test the control is half the traffic. In a bandit or a long program, 10% to 20% kept on the original is common.
How Luria uses one
Your unchanged page is always the control, never below 10% of traffic. Each visitor sticks to one version. Luria's 60-day guarantee is measured against that control: if no version beats it at 95% certainty (with 5,000+ tracked visits on the tested pages), that period's fees come back. Full rules.
Questions
Is a holdout the same as a control group?
Yes, in page testing the words mean the same: visitors who see the original. Holdout usually means the share stays on the original for the whole program, not just one test.