Consumer behavior is of quintessential importance to the retail industry. Purchase trends for retail products are often affected by consumer location. Region-specific purchase trends, such as the French love wine, are typically supported by anecdotal evidence. An automated technique for the discovery of such trends from retail data has so far been absent. In this paper, we address the challenge of colloquial region discovery for retail products. More specifically, we target the problem of examining product sales across a chain of stores to extract the geographic regions that characterize a product. We introduce DICE, a diffusion-based technique to uncover all such regions for a given product, when they exist. In contrast to the current state of the art, DICE involves minimal usage of parameters and shows remarkable tolerance to noise that is often ubiquitous in retail data. We present results of experiments conducted on real datasets from a supermarket chain in France. Empirical evaluation and user studies establish that the proposed technique significantly outperforms the natural baseline and previous state-of-the-art approaches. Further, we study the impact of time and product category on DICE and discuss use cases for application of DICE in the retail world.