RCTs for QSR Chains

Most organizations now embrace experimentation, conducting A/B testing, the gold standard of which is a randomized control (RCT). An RCT introduces one intervention to a randomly selected group of participants, then compares the results to a control of "business as usual ."

RCTs are still challenging for large, franchise-dominant QSR chains, which have contractual obligations on tv buys that make it difficult to conduct clean RCTs. Plus, they must coordinate equipment, ingredients, and training for a product that may only get created and sold for the select group of customers receiving the test intervention. Yet market tests are still filled with "noise" that's difficult to control.

One solution is to soft launch a new product after mitigating all other risks through research and operational pressure tests. Then conduct RCTs on messaging in a digital channel. For example, Taco Bell could test messaging through its kiosks, now in most of its stores. If the new message performs poorly, Taco Bell can avoid misspending its expensive ad budget and sending its heavy users to a competitor with a more relevant message.