Experimental Design Flashcards
Recipe
Uniue combination of variables
Also known as treatment
A-B Split Testing
A - original recipe
B - challenger recipe
Typically few designs (up to 10 at a time)
Sequential tests less reliable since changes might be due to other factors
A-B Split Testing Advantages
Easy to test, implement and analyse
Flexibility defining variable/factor values (revolutionary combined with evolutionary designs)
A-B Split Testing Disadvantages
Not possible to test all possible combinations/find interdependence between all factors/variables so not likely to find optimal combination
Can’t see which factors are contributing the most as too many things change simultaneously
Inefficient form of data collection
OFAT
Often used but inefficient