Experiments to evaluate alternative collection methods are rare for establishment surveys. In support of a re-engineering of the 2017 Economic Census, the United States Census Bureau began conducting experiments to evaluate alternative contact strategies. The goals of this effort are to improve the timeliness of response as well as overall response, and to reduce collection costs by reducing the number of nonrespondents re-contacted by more expensive follow-up methods. As the Economic Census is conducted every five years, a series of experiments has been implemented using the Census Bureau’s annual and quarterly establishment surveys to test the effectiveness of less-expensive mail strategies. The first of these experiments was incorporated into the Business and Professional Classification Survey (SQ-CLASS), a web-based survey conducted quarterly for the purpose of matriculating “births” (businesses with new or recently reactivated Employer Identification Numbers) into the Census Bureau’s business survey frame, and assigning them industry classifications. We conducted an experiment in the third-quarter 2014 SQ-CLASS to test the effectiveness of sending a reminder letter prior to the survey due date. About five thousand nonrespondents were randomly selected to receive a reminder letter about two weeks before the due date, and the others received only the typical post-due-date follow-up. Nonrespondents in the experimental group also received the usual post-due-date follow-up. The experimental group’s cumulative response increased relative to the control group following the mailout of the pre-due-date reminder. In this presentation, we will describe the design of the experiment and the methods of analysis, and discuss the results in the context of similar research with demographic surveys and theories about survey response.