How big should the holdout be?

10% is the standard for most teams. 5% is the minimum for strong-lift detection. 15% is appropriate for small audiences. Above 15% is rare and usually unjustified.

How long should the holdout window be?

90 to 180 days for meeting and qualified opportunity metrics. 6 to 9 months for closed-won metrics. The window must be long enough to capture the full conversion cycle.

Can holdout prospects be touched later?

After the conversion window closes, yes. The cohort can be re-included in normal cadences for the next quarter. During the window, no, or the data is contaminated.

What if the lift is small or null?

Important to know. A small or null lift means the cadence is not adding much over organic conversion, which usually points to a targeting problem or a cadence quality problem.

The Holdout Group Method: How to Prove Multichannel Lift

A 5 to 10% holdout group proves multichannel lift is causal rather than correlational. The conversion gap between holdout and treatment is the actual cadence-driven lift on top of organic conversion.

Holdout Design

Random assignment within persona, persistent assignment for 90 to 180 days, hard suppression on cadence platform, track conversion for both groups, compute lift weekly, refresh quarterly.

Holdout Size

5% minimum, 10% standard, 15% for small audiences (1,000-3,000 prospects). Above 15% the opportunity cost outweighs precision benefit.

Metrics To Measure

Conversion to meetings booked, qualified opportunity, closed-won. Average deal size by group. Time to convert. Report all 5 across treatment vs holdout.

Common Objections

Insufficient volume (use program-level holdouts), leaving meetings on table (precision pays back 5-10x), ABM teams (program-level holdouts), SDR overrides (hard suppression), leadership pushback (frame as measurement investment).