Fractional factorial designs are something that is common in the field of industrial statistics, where people seek to test out and evaluate different aspects of product development. Easier to use than the traditional factorial methods, they allow you to work on a small and manageable scale to conduct statistical experiments. This allows greater flexibility when it comes to forecasting for your product, thereby allowing for enhanced decision-making.

Like every statistical approach, there are some constraints to using traditional factorial designs. The main one is that if you throw too many variables into the mix, you can end up with an infinite number of permutations, and create a matrix that is impossible to work with. Of course, all good statistical procedures rely on carefully and thoughtfully specified model parameters, but this can nonetheless be problematic.

The Benefit of Fractional Factorial Designs
A benefit of fractional factorial designs is that they allow you to address this issue. You can still produce a decent estimate of main effects within your model, while keeping it smaller and more manageable than its full-scale counterpart. This in turn allows greater flexibility with your analyses.

If you are interested in using statistics to create a better product or to help you maximize the production and marketing process, your best bet is to consult a trained statistician. Whether for DOE, DFSS, or factorial work, he or she can talk through your project and guide you to the best analytic approaches. Critically, a consultant can provide assistance with interpreting the output, which can be a daunting task for those who lack training.