I sometimes get called up to fix issues with Excel running too slow or throwing cryptic error messages that not only confound my local user base, but, in checking, the entire Googlesphere of users which never arrives at a consistent definite resolution. The rate of resolving these issues by simply manhandling the local PC or instance of MS Office is something like zero percent.
If an Excel instance is generally running fine, as is the OS as a whole, more often than not I wind up with the sense that whatever is being "Excel-processed" probably shouldn't be, or, if ever should have been and was, is beginning to reach its implied level of non-suitability for the task. It's either beginning to work with a dataset that is increasingly too large for stability, or there has been an increase in the inter-dependencies of the specific worksheet as a whole -- usually both. The issues sometimes affect just one user, a small subset, or everyone using the same "data solution" called Excel.
It's not my place to play role of "data analyst" in these situations, but, just in knowing how these things work over 20 or so years, my guess is always that an "Excel-centered solution" that at a lower scale was just fine, is being applied without check or re-analysis, as the larger scale use case begins to unfold.
Direct link to video if it doesn't play right here, is here instead.
By Dave for Personal Blog.