Let me first quote the definition of Ricci curvature from Wikipedia. Let $M$ be an n-dimensional Riemannian manifold equipped with its Levi-Civita connection $\nabla$. The Riemannian curvature tensor $R$ of $M$ is the $(1,3)$ tensor defined by $$ R(X,Y)Z=\nabla_X\nabla_YZ-\nabla_Y\nabla_XZ-\nabla_{[X,Y]}Z $$ on vector fields $X,Y,Z$. Let $T_pM$ denote the tangent space of $M$ at a point $p$. For any pair $(\xi,\eta)$ of tangent vectors at $p$, the Ricci tensor $R(\xi,\eta)$ evaluated at $(\xi,\eta)$ is defined to be the trace of the linear map $T_pM\rightarrow T_pM$ given by $$ \zeta \mapsto R(\zeta,\eta)\xi. $$
My question is why is this a reasonable definition? It seems to me that there is more natural linear map $T_pM\rightarrow T_pM$ given by $$ \zeta \mapsto R(\xi,\eta)\zeta $$ and we may take the trace of this map. In this case what shall we get?