Well, you've got some problems, but it's hard to understand without a chromatogram. Here's some tips.
You'll have to determine dead time, it can be no later (and often a little earlier) than the time it takes your particular flow rate to sweep through an empty column -- pi*radius in cm*length in cm. Dead time often, but not always shows up as a small peak like you observed, it's caused by the flow disturbance caused by the injection.
Try to build a table, of dead time (we call that Vo), and your peaks, and their retention times. If you feel like it, compute average, and std dev of the retention times, so you can clearly see, how good or bad your results are. Maybe, a really bad outlier, will prompt you to realize, "Wait, I've been misidentifying all along."