Combinatorial Design

When to use

"Use when the solution space is a combination of multiple options across multiple dimensions. Systematically explore combinations to find optimal or near-optimal configurations, rather than testing a few intuitive guesses."

Example

"A/B testing is combinatorial design at its simplest — test two variants and pick the winner. Multi-variate testing extends this — test combinations of headline, image, CTA, and layout to find the optimal combination."

Common misuse

"Exploring combinations exhaustively when the space is too large. Combinatorial explosion makes exhaustive search infeasible; use heuristics, sampling, or optimization algorithms instead."

Also known as

Related models