Public health officials enforced several measures to contain the COVID-19 pandemic that affected grocery stores, such as limits on store capacities and enforcement of masks and physical distancing among customers. Nevertheless, these measures can provoke queues, which could drive customers away from stores. In this study, we investigate how customers trade off between social distancing measures and increased waiting times during the peak of the COVID-19 pandemic. Our data comes from an online survey applied in New York City in May 2020. This survey included a set of discrete choice experiments framed in virtual stores, as well as a set of psychometric indicators regarding the pandemic. With this data, we estimated a latent class conditional logit model where assignment to classes is correlated with COVID-19 latent variables. We identified three latent classes with preference structures that valued social distancing to varying degrees. In spite of this heterogeneity in preferences, we found that customers were willing to wait longer to access stores with better social distancing measures. This result suggests that stores could increase, rather than decrease, their sales if they enforce public health measures at the expense of longer waiting times.