In this paper, we revisit the semantics of recommendations and promotional offers using multi-objective optimization principles. We investigate two formulations of product recommendation that go beyond traditional settings by optimizing simultaneously two conflicting objectives: Budget-Reco optimizes two customer-centric goals, namely utility and budget, and Business-Reco optimizes utility, a customer-centric goal, and profit margin, a business-oriented goal. To capture those objectives, we formulate knapsack problems and propose adaptations of exact and approximate algorithms. We also propose Group-Promo, the problem of generating product promotions that we model as a group discovery problem with multiple objectives and develop a Pareto-based solution. Our experiments on our TOTAL datasets demonstrate the importance of multi-objective optimization in the retail context, as well as the usefulness of our solutions when compared to their exact baselines. The results are valuable to TOTAL's marketing department that has been improving hand-crafted strategies by launching several promotional campaigns using our algorithms.