I've worked for a massive corporation in which they had 300mb+ excel files they bought high specs computers just to have them load fast enough and searching would take 3 to 5 minutes we suggested that they'd try moving it to Microsoft access and the query became instant, I can't imagine the hours wasted waiting for the queries to run
I used to manage a team that developed tools for traders, in excel. We had excel sheets that loaded 20+ custom addins, most related to financial libraries. The excels would compute prices of financial products in real time, and send those prices live on multiple platforms. Those prices were executable, meaning other clients and banks could trade live in a click.
This was (probably still is for sken banks) the norm 10 years ago. Most banks were running their whole trading desks out of excel.
The data fits on the screen. Roughly 40x30 matrices are fine. These rules aren’t set in stone, so there are situations when it’s still justified to use matrices larger than that.
You can see all the sheets you need. Roughly 10-14 sheets. If you need more sheets than that, you should probably split the calculation into several files.
You need complicated VBA macros and you need them to run perfectly every time. In my experience, this programming language is infuriatingly unpredictable and unreliable. Random things happen all the time and no amount of debugging is able to solve these types of cosmic problems.
You can go beyond those general guidelines, but using everything gets more and more annoying the further you go. Pretty soon you’ll realize it would have been better to build the whole thing in Python, R or something else. Once the file size hits 15 MB you know you’ve gone way too far and it’s about time you rebuilt the whole calculation using some nicer tool. I try to switch as soon as possible when I realize my calculations are about to grow beyond these limits.