Purpose – The purpose of this paper is to address problems that exist in the context of XML to ontology translation. Existing research results dealing with XML to ontology translation do not facilitate bidirectional data translation due to the fundamental differences between XML schema and ontologies. This paper proposes a mapping representation ontology for modeling concept mappings defined between XML schema and ontology, enabling data translation without any information loss. Design/methodology/approach – A two-step compensation approach is proposed that aims to prevent the loss of data type, structural and relational information during any single trip data translation. The mapping representation ontology proposed is capable in capturing enough information to compensate the loss of information during translation, hence allowing bidirectional conversions between XML and ontology. Findings – Fundamental differences between XML schema and ontology are identified as the main reason causing the loss of information during data translation. A compensation approach that captures a sufficient amount of concept mapping information data translation is found to be successful in enabling lossless data transformation. Practical implications – Outcomes from this work allow for the seamless data translation between XML documents, it demonstrates how web applications can seamlessly communicate and exchange data with each other without the need to conform to a predefined data standard. This paper aims to enhance interoperability between distributed systems. Originality/value – This paper presents a mapping ontology that captures concept mappings defined between XML schema and ontology. Two algorithms facilitating the bidirectional XML to ontology translation are also proposed.