Can someone please guide the how would be the graph of no. of users Vs response time. I want to know that what is general standard for degradation in response time when no. of user increases in the application. For example, there is 10% degradation in response time when user increases from 1 to 25 and so on.
A highly advanced bug is indistinguishable from a feature.
Try plotting no. of users versus response time for a sample set wide enough to make meaningful assertions. Increasing users from 1 to merely 25 may not lead you to meaningful result. In general you can try plotting degradation against a standard exponential decay. If you are not much concerned about very finer details then just see whether your degradation pattern is slower than exponential decay or otherwise.
You should worry if degradation is worst than exponential decay. If you are keen to know more about degradation patterns then try putting your hands on sampling techniques intricacies. This board may not cater the need for elaboration of intricacies why we should try exponential decay as a preliminary benchmark for comparison. Infact that is a suggestion for you to start with, not an Industry mandate yet.
There is a good example of this graph in Steve Splaine's Web Testing Handbook and in my article on creating the curve www.perftestplus.com/articles/userexp10.pdf This explains the different areas of the curve and the general shape and pattern.