Ontologies have become ubiquitous tools to embed semantics into content and applications on the semantic web. They are used to define concepts in a domain and allow us to reach at a common understanding on subjects of interest. Ontologies cover wide range of topics enabling both humans and machines to understand meanings and to reason in different contexts. They cover topics such as semantic web, artificial intelligence, information retrieval, machine translation, software development, content management, etc. We use ontologies for semantic annotation of content to facilitate understandability of the content by humans and machines. However, building ontology and annotations is often a manual process which is error prone and time consuming. Ontologies and ontology-driven content management systems (OCMS) evolve due to a change in conceptualization, the representation or the specification of the domain knowledge. These changes are often immense and frequent. Implementing the changes and adapting the OCMS accordingly require a huge effort. This is due to complex impacts of the changes on the ontologies, the content and dependent applications. Thus, evolving the OCMS with minimum and predictable impacts is among the top priorities of evolution in OCMS. We approach the problem of evolution by proposing a framework which clearly represents the interactions of the components of an OCMS. We proposed a layered OCMS framework which contains an ontology layer, content layer and annotation layer. Further, we propose a novel approach for analysing impacts of change operations. Impacts of atomic change operations are assigned individually by analysing the target entity and all the other entities that are structurally or semantically dependent on it. Impacts of composite change operations are analysed following three stage process. We use impact cancellation, impact balancing and impact transformation to analyse the impacts when two or more atomic changes are executed as part of a composite or domain specific change operation. We build a model which estimates the impacts of a complete change operation enabling the ontology engineer to specify the weight associated with each optimization criteria. Finally, the model identifies the implementation strategy with minimum cost of evolution. We evaluate our system by building a prototype as a proof of concept and find out encouraging results.