In two studies, we investigated the impact of instructors' different knowledge bases on the quality of their instructional explanations. In Study 1, we asked 20 mathematics teachers (with high pedagogical content knowledge, but lower content knowledge) and 15 mathematicians (with lower pedagogical content knowledge, but high content knowledge) to provide an explanation about an extremum problem for students. We found that the explanations by teachers and mathematicians mainly differed in their process-orientation. Whereas the teachers mainly presented the solution steps for the problem (product-orientation), the mathematicians also provided information to clarify why a certain step in the solution was required (process-orientation). In Study 2, we investigated the effectiveness of these differing explanations. Eighty students either received a process-oriented mathematician's explanation, a product-oriented mathematics teacher's explanation, or no explanation for learning. We found that students who learned with a process-oriented explanation outperformed students who learned with a product-oriented explanation on an application test. Students who only had the problem but no explanation for learning showed the lowest learning gains. Apparently, deep content knowledge helped instructors generate explanations with high process-orientation, a textual feature that served as a valuable scaffold for students' understanding of mathematical procedures.