Libre calc is very lacking, compared to Excel. Same with Base - Access. No Outlook alternative at all. No cloud integration. With office365 you get Sharepoint AND Exchange servers.
Most people don't use Access.
Joe Marketing ain't going to jump through UI hoops and do complex bug work-arounds, when he can avoid this entirely by buying a MS office license for 200 USD or whatever the price is.
A couple of the senior people write queries and import routines, and the clerks/etc use them in their workflow. Works great.
Also, Calc has Python (and other language scripting) which puts it above Excel in my book. For built in functions yes Excel is better. As soon as you start venturing outside of Spreadsheets and into data analysis (where Macro's come into play) stop using Excel. Move to Matlab or Pandas. Excel Macro's should not be used as a selling point, they are rarely, if ever, the right tool for the job. At least with Calc you can take the business logic from your Macro's and move them into a real script. VBA isn't C# nor VB.Net compatable.
I see Excel users in 2 camps:
Those that don't know what functions are, and if the do, tread carefully with '=if(...,...)' and get intrigued. Though those that get intrigued are in a minority.
Those that know what functions are have Excel act as a full-fledged IDE, which it does pretty well. This second camp either write libraries to call from Excel, or have someone write them in C++ if the libraries are that important.
Excel is exceedingly good for hacking, or I should say, playing. It is not a spreadsheet, it is an extension of skin from keyboard to screen for visualizing data, from brain to 2D hackable multi-dimensional grid, that is as easy as playdough.
Vendors include SAS, Oracle, IBM, Predixion.
Also Excel is used as a self-service BI tool. It has ETL, OLAP and dynamic dashobarding.
Excel (and other spreadsheet packages like Calc) are good for non technical users it's useful for rapidly prototyping stuff but it is absolutely a pain for automation and any type of statistical heavy lifting. Once a spreadsheet reaches a certain level of complexity it becomes a pain to support and maintain much better to have a real database driving things.
I've heard a lot about R which is kind of an open-source version of SAS (at least the statistical parts) not sure what its automation and reporting support is like my limited understanding is it doesn't have the BI parts.
If Excel had a real scripting language as it's Macro language the problem wouldn't be nearly so bad. All that business code could just be copy/pasted, have some unit tests implemented then refactored so that the accountants code is now updated by the developer. As it is now the code has to completely rewritten in a new language.
As for the language choice, VB.Net or C# would be fine. I'm not saying it has to be Python. My complaint is that the code is non-transferrable. The apps cannot grow past Excel, and the email the most recent version to the next guy game gets you into versioning hell rather quickly.
He wants his pivot tables, KPIs and Solver and some macros. He wants point-and-click ETL. If it's a department/cross-department model, do it in SQL Server, or whatever else you use. You can use your IT kung-fu there. This is about what libre office lacks. Excel has power and flexibility and remains easy to use for non-programmers.
My issues with Excel come from the perspective of modifying these ad-hoc apps (many business run off these for a long time as they grow). Yes, there are other great features, but for me, for growth, Calc is taking a better approach.