This sort of experiment, just like any other experiment, requires a control! There are so many things controls tell you, and the risk of getting a false finding certainly justify the extra effort/money put into making the control experiment.
In immunofluorescence, this is especially important whether being done for research or for disease diagnostic purposes. The best reason I can think of is the possibility of cross-reactivity or side reactions with your antibodies.
Often, we discover cross-reactivity or other issues completely accidentally! Think about how many poxvirus antibodies are cross-reactive (basis for Variola vaccine? Vaccinia!), or how Epstein-Barr Virus allows the formation of heterophile antibody that will agglutinate horse and sheep blood! You may end up with a weird reaction, but without a control, you'd never know that a certain result of the experiment is caused by your changing variable or by some other extrinsic factor!
The solution? Make a control! Do an experiment with your equipment where you already know what the result will be (or rather what you expect the result to be based on established empirical data). Then you can report that your control gave expected results, and that the results based on your changing variable can only be the result of your changing variable. Keep in mind, too, there are often multiple controls in a given experiment to provide statistical significance.
I hope that helps with your problem!
I worked in the Department of Research and Development in a Biotech company. For this question, I discussed (or argued) with the department manager since 2008, because I believed that there was no controls performed parallel to experiments in the research using the method of immunofluorescent staining. I believed that results obtained from those experiments were confirmation bias. In general, images of those experimental results were art works rather than scientific results. Many those images were published in scientific journals. I had paid a great cost for this disagreement.