At the beginning of every test, our experts get more familiar with the server efficiency expectations and define the requirements. After analyzing the functions and infrastructure of the system, testers review typical user behavior (by interviewing the system owner, checking data from Web Analytics tools, or application logs). On this basis, our team builds a load model, which contains information about user actions and their visit on the site, as well as the percentage of users performing the action. After running tests, we indicate locations that are responsible for insufficient system performance. The test ends with a detailed report, in which we describe potential causes for low performance.