CRM glossary
Holdout group
Also called: holdout, control group, suppressed control
A holdout is a randomly chosen slice of the eligible audience deliberately excluded from a campaign or programme, so its behaviour shows what would have happened anyway. Comparing treated contacts against held-out ones turns an open rate into a revenue claim you can defend in a board meeting. Keep the selection random rather than convenient, size it so a revenue difference is detectable, and run a long-lived programme-level holdout alongside campaign ones — the campaign view hides cannibalisation between messages.
Articles using this term · 11
- A testing cadence that survives a busy quarter · 18 August 2026
- Connecting CRM activity to revenue you can defend · 28 July 2026
- Marketing mix modelling for CRM teams · 11 March 2026
- The decision log that stops two years of tests evaporating · 18 December 2025
- A loyalty programme that survives a margin review · 5 November 2025
- Peak season: what we changed after last year · 2 October 2025
- Propensity models a CRM team can actually run · 13 September 2025
- Holdout groups, or you cannot prove CRM works · 5 June 2025
- A CRM year in review that finance will read · 20 December 2024
- Basket recovery that doesn't train people to wait · 22 November 2024
- Life after the open rate · 22 June 2024