It rarely means the test is wrong. It usually means the reviewer got to the end of the paragraph and still could not tell what was compared with what, how many times, and in how many animals, cells or runs.

What they are actually asking

Three questions, in this order:

• What is n, and what is it counting? Three technical replicates of one sample is not three samples. Say which it is, in the caption, every time.
• What was compared? Two groups, or six groups with one control? If six, the test has to know that.
• What does the error bar show? Standard deviation, standard error, or a confidence interval — they differ by a factor that decides whether the reader believes you.

A caption that answers all three usually makes the comment disappear, without a single number changing.

The part people get wrong

Running a two-sample test on every pair in a six-group experiment gives you fifteen tests and a very good chance that one of them looks significant on its own. Reviewers who have seen that before will assume it until you show otherwise. Either correct for it, or state plainly that the comparisons are exploratory and not corrected — that sentence is a defence, and a p-value quietly left uncorrected is not.

Write the analysis before you collect the data

Not for the reviewer — for you. If you cannot say in advance which test answers your question, the experiment is not designed to answer it yet, and no test chosen afterwards will fix that.