Submodular function maximization is central in numerous data science applications, including data summarization, influence maximization, and recommendation. In many of these problems, our goal is to find a solution that maximizes the \emph{average} of the utilities for all users, each measured by a monotone submodular function. When the population of users is composed of several demographic groups, another critical problem is whether the utility is fairly distributed across groups. In the context of submodular optimization, we seek to improve the welfare of the \emph{least well-off} group, i.e., to maximize the minimum utility for any group, to ensure fairness. Although the \emph{utility} and \emph{fairness} objectives are both desirable, they might contradict each other, and, to our knowledge, little attention has been paid to optimizing them jointly. In this paper, we propose a novel problem called \emph{Bicriteria Submodular Maximization} (BSM) to strike a balance between utility and fairness. Specifically, it requires finding a fixed-size solution to maximize the utility function, subject to the value of the fairness function not being below a threshold. Since BSM is inapproximable within any constant factor in general, we propose efficient data-dependent approximation algorithms for BSM by converting it into other submodular optimization problems and utilizing existing algorithms for the converted problems to obtain solutions to BSM. Using real-world and synthetic datasets, we showcase applications of our framework in three submodular maximization problems, namely maximum coverage, influence maximization, and facility location.