Growing enrollment in distance education has increased student-to-lecturer ratios and, therefore, increased the workload of the lecturer. This growing enrollment has resulted in mounting efforts to develop automatic grading systems in an effort to reduce this workload. While research in the design and development of automatic grading systems has a long history in computer education, only a few attempts have been made to automatically assess spreadsheet and database skills. This paper has three purposes: (1) to describe the design of an assessment in the "Information Systems" course at the Open Polytechnic to assess students' spreadsheet and database skills, (2) to describe the development of an automatic grading system to assess spreadsheet and database skills, and (3) to compare automatic with manual marking to determine if automatic grading system is a feasible method of reducing workload. The automatic grading system we developed uses Excel's user-defined functions to automatically check whether a feature or a function has been used. Since the outcomes from user-defined functions are scrambled, students verify their own answers by entering the results from these functions into an online quiz. As a result, there is no need for the lecturer to download, open, and check the actual software application. The system recognizes correct answers from these scrambled inputs and allocates marks. This system is integrated into the Moodle learning management platform and linked to the students' academic record database. The main difference between the automated grading system for the assessment of spreadsheet and database skills described in this paper and existing systems is that the latter systems require the actual software application to be submitted for marking. The system described in this paper does not require markers to handle the application. Instead, it automatically checks the application while students are working on it, but grading is not performed until students answer specific quiz questions. Practical experience with the automatic grading system has shown that the system significantly decreases turnaround time for the grading of assignments, while providing instant feedback to students on the correctness of their answers. At the same time, the system reduces the workload of the lecturer, freeing lecturers from administration and the time-consuming tasks of checking individual aspects of the spreadsheet and database applications. This allows them to allocate time to student support and other more creative activities. In addition, the automatic grading system allows for a much finer probing of individual aspects of the spreadsheet and database applications, with no additional work required by student or lecturer. The methods of marking were evaluated to address the main research question of whether there were significant differences between a human and an automated grading system. A comparison between the methods of marking (human and the automatic grader) based on data from 11 trimesters indicated no significant difference in the average marks and mark distributions in the case of the spreadsheet application. The comparison also showed that although the difference in the average marks in the case of the database application was significant, it did not mean that the effect of the method of marking was meaningful or important, as illustrated with the effect size. Nonetheless, monitoring of the automatic grader results is recommended. (Contains 6 figures and 2 tables.)