A capable person, armed with a clear goal and a willingness to learn, hits a wall built by the tool itself. The user spent 90 minutes wrestling with forum solutions, then faced the grim prospect of 28 hours of manual data entry. That math doesn't add up. The spreadsheet should be the solution, not the problem.
What's happening here is a fundamental mismatch between how the data arrives and how the human needs to use it. Mentimeter exports responses horizontally, columns for respondents, rows for answers. The user wants a vertical layout: questions in column A, answers in adjacent columns. This is a simple transpose operation in concept, but in practice it requires chaining INDEX, TRANSPOSE, and careful absolute references. The autofill logic fights you because you're asking Excel to move across columns when it expects to move down rows. The user tried the right formulas. The tool just wasn't designed to make this easy.
We think this story reveals something important: the spreadsheet model itself is the bottleneck. The grid treats rows and columns as equally valid containers, but your brain works in lists and hierarchies. When you export survey data, you don't want to think about cell references. You want to see respondent 1's answers grouped, then respondent 2's. The user's request, put each respondent on their own sheet, flip the axes, isn't exotic. It's the most natural way to review 20+ people's feedback. That Excel makes it a 28-hour problem is a design failure, not a user failure.
An AI-native spreadsheet would recognize the pattern in the Mentimeter export and ask: do you want this transposed? Do you want each respondent in a separate sheet? It would execute the transformation in seconds, not hours. The user shouldn't need to be an Excel wizard to view survey results. The tool should understand the intent and offer the right structure. That future is accessible now. The only thing standing between this user and their 28 hours is a tool that treats data entry as a puzzle rather than a workflow.