One approach to this is to set up before you start an indifference zone which says something like "I am not interested if the difference is less than +/-5 mm." Then find a 95% confidence interval for the difference using the data. Plot both the indifference zone and the CI on a number line and inspect the diagram.
Now a variety of conclusions are possible such as "the difference is real and important" or "the difference is real but not important" or "the difference is real but we can't say if it is important or not" or even "the difference may be important but we can't even say if it is real." It depends on how theCI overlaps the indifference zone and whether it contains 0.