When variables are plotted on a graph, it is common for the values to cluster together. When you have one dot that it all by itself far, far from the other data points, we called that an outlier.
On a scatter graph plotting the interest rates being paid by various bonds, the data points all clustered together. One dot showed an interest rate far above the others. This dot represented a company that was tottering on bankruptcy. It was an outlier, paying much higher interest to compensate for the increased risk.