Several different methods are used for assessing the usability of websites. The 'usability' of a website describes how competently and easily a user or visitor without any proper training can interact with the site. Nowadays, social networking systems have become a major medium for online communication, marketing, research and collaboration. Users can quickly be in touch with friends and colleagues via a social networking site.The aim of this paper is to investigate the usability of social networking systems, using expert evaluation and user testing, in order to provide recommendations for developers in the improvement of social networking in general and LinkedIn in particular, based on the difficulties (usability problems) experienced by users and experts, and also to measure the users' performance (both novices and experienced users). In this experiment, the researchers employed two usability evaluation methods: 'user testing' and 'expert evaluation'. The 'user testing' includes pre-test, tasks to be performed and post-test usability ratings. The 'user testing' is conducted using a thinkaloud protocol and observation to achieve a better understanding of the participants' behaviours. The 'expert evaluation' includes pre-evaluation, actual evaluation (each of the ten Nielson's heuristics was broken down into a set of principles forming a checklist specifically oriented for the LinkedIn website), followed by post-evaluation.The findings of this research highlighted a number of usability problems, but none of them is catastrophic. Some of the usability problems discovered were: 1) confusing or unfamiliar terminology such as 'friends' being referred to as 'connections'; and 2) the unexpected location of components or links. It was also found that there was some variations in opinion between novices and experienced users regarding the user interface usability of LinkedIn, as the novices had difficulty in performing some of the tasks.It seems that the LinkedIn website was highly usable, the evidence being that 61% of the problems discovered during the expert evaluation were cosmetic, which means that they could be fixed if enough time was available. The difficulties discovered by the novice users were 35% minor and 4% major, and they experienced no catastrophic usability problems. A set of recommendations for developers in the improvement of the social networking are provided at the end of this experiment.