The prediction that the effect of geographic distance as a barrier for trade will decrease over time and eventually be eliminated as a result of technological developments has been proposed for decades. However, empirical evidence suggests the negative effect of geographic distance persists in both international and domestic trade. This study uses data from a national survey of nursery and greenhouse growers in the United States to investigate the effect of geographic distance on trade. We model trades as a two-step process in which selling firms and buyers were matched first, and then a positive trade flow occurs conditional on the match. This process leads to a simple zero-inflated estimation strategy and provides a meaningful interpretation of the zero-trades. We demonstrate that this approach predicts the overall trade flow better than the Poisson or Gamma Pseudo Maximum Likelihood (PML) approach, due to its ability to predict zero trade flows more accurately. The empirical results indicate the effect of distance on trade volume is large and statistically significant. In addition, firms trade almost exclusively with partners in the home state or in states with branch offices. We find that social media and website advertising is much less effective in increasing the trade flow than traditional advertising channels.