The purpose of this study is to analyze two estimation models related to relationship marketing in business-to-business interactions in the context of emerging economies. We compare two estimation models – one based on a dyadic approach and another based on a non-dyadic approach. We estimate these two models and compare their results to see which one is more suitable to be used as a theoretical relationship marketing model. We developed a survey comprising 204 dyadic observations of retailers and their suppliers and used a purposive sampling method. Three different observations correspond to three different estimation models. This study shows that the estimation model based on a dyadic approach has a better model fit than the model based on a non-dyadic approach regarding relationship marketing in a business-to-business context. The dyadic model also gives more accurate information to explain behaviour among companies involved in business relationships. Furthermore, this study explains how to examine estimation models in a relationship marketing context using dyadic and non- dyadic approaches. We also develop methods for examining dyadic perceptions of companies involved in business relationships. Our results contribute to the IMP school of thought in relationship marketing.