Online Excel test to assess candidates

Through two complementary mechanisms. The first is the knowledge test, of the Software Skills type, with auto-graded multiple-choice questions on formulas, functions, pivot tables, and best practices, ideal for screening at volume. The second is the hands-on task, through the File question type: the candidate receives a dataset and a real request (clean the data, put together an analysis, build a dashboard), solves it in the Excel or Google Sheets they use every day, and uploads the finished file, which the evaluator opens and reviews against the defined criteria. Knowledge tells you whether the person knows; the task shows how they work.

Because a simulator tests how well someone adapts to the simulator, while the task tests the work. Simulated spreadsheet environments behave differently from real Excel: shortcuts that don't work, missing functions, odd performance, and part of the score ends up measuring how well the candidate adapted to the test tool. In the file task, the person uses the software they'll use on the job, with the shortcuts and features they know, and delivers a result the evaluator examines the way they would examine a coworker's work: formula structure, sheet organization, data handling, clarity of the result. For anyone hiring an analyst, it's the most accurate picture available.

The company's evaluator, guided by the criteria defined when the question was created. When setting up a File question, filling in the goal and the evaluation method is mandatory: what a good submission contains (correct formulas, clean data, an analysis that answers the question) and how to score it. This makes two different evaluators look at the same spreadsheet with the same yardstick. The objective part, the quiz, is graded automatically on the spot, so in a combined assessment the recruiter already gets the knowledge score ready and only manually reviews the hands-on submissions.

The company's evaluator, guided by the criteria defined when the question was created. When setting up a File question, filling in the goal and the evaluation method is mandatory: what a good submission contains (correct formulas, clean data, an analysis that answers the question) and how to score it. This makes two different evaluators look at the same spreadsheet with the same yardstick. The objective part, the quiz, is graded automatically on the spot, so in a combined assessment the recruiter already gets the knowledge score ready and only manually reviews the hands-on submissions.

They can try, and Coodesh's answer is the same as for technical tests: make it visible and make it pointless. Visible, because the assessment's integrity features log tab switching and response time, and the pattern of someone who outsourced the task shows up in the review (formulas too sophisticated for the rest of the submission, an inconsistent approach). Pointless, because the most effective protection is to add a video question or an AI interview to the same assessment asking the candidate to explain their own spreadsheet: why this function, what this pivot table solves, what would break if the dataset doubled. Whoever did it can explain it; whoever copied it gets stuck.

For any role in which the spreadsheet is a core work tool: financial, operations, HR, and marketing analysts, controllership, planning, customer success teams that handle reports, and administrative assistants. Calibration matters more than the test: an assistant needs VLOOKUP, filters, and formatting; a financial analyst needs pivot tables, nested functions, and modeling; a data analyst may need SQL more than Excel, and Coodesh has both in the same library. Setting the level based on how the role actually uses spreadsheets keeps you from rejecting good people over skills the job doesn't require.

Yes, and that's where the format pays off the most. In My Library, you create a File question with a dataset similar to the one you use every day (anonymized, with no real customer data), describe the task as you would to a new coworker, and set the grading criteria. The question is exclusive to your company, isn't added to the public catalog, and can be combined with the library's Excel quiz and a video question in the same assessment. The right candidate recognizes the problem as real, which also improves their experience with the process.

Within the overall assessment result: the quiz score comes in automatically, the hands-on task comes in as soon as the evaluator reviews it, and the final score is the weighted average based on the weight you gave each part. The uploaded file is available for download during review, performance is broken down by skill, and, if integrity features are on, a summary of suspicious activity comes with the result. Everything can flow back to your ATS through the API integration.