The Markowitz mean–variance-optimization framework presents a puzzle. Although it is the standard model of portfolio construction, investors rarely use it and, when used, it is constrained so much that portfolios reflect the constraints more than the optimization. Why do investors dislike unadulterated optimized mean–variance portfolios? The typical answer focuses on biases introduced by estimation errors. More important is the fact that investors' goals are quite different from mean–variance optimization. People care about more than cost and nutrition when they construct their diets: They also care about palatability. Similarly, investors care about more than expected returns and variance as they construct their securities portfolios: They also care about the investment equivalent of palatability. We use an analogy between security portfolios and food portfolios to explore the nature of the gap between intuitively appealing portfolios and mean–variance-optimized portfolios.