I have worked for companies with 200+ developers, in which case I have not had a choice what method to use.
I have worked for companies with 3 developers and zero IT in which case I am the IT support and have had to make sure source code gets handled in a good way.
I have worked for a number of years on my own, in which case I still have huge advantages for repositories.

I have worked with projects with million+ lines of code.
I have worked with projects with less than 10k lines of code.

I have worked with projects saving to zip files.
I have worked with projects duplicating directories.
I have worked with projects using real version-control software.

I have been running a CVSNT repository on my own machine.
I have been running a virtual machine as repository server.
I have set up dedicated physical machines for source code repository.
I have worked in organisations where dedicated repository machines have already existed.

I have basically tried all combinatins.
The repository route is always better than any copy/compress method.
No extra hardware is needed, sw had for free and the initial setup is negligible.
The only extra time involved during project is to write comments when checking in changes - and that time tends to save time down-the-line. This step can be skipped if people don't believe in the value of checkin comments.

ZIP files still needs backup.
Duplicated directories still needs backup.
A CVS or SVN repository is just a directory tree that is backed up exactly like other directories, so no extra time for setting up any backups unless the repository is hosted on a dedicated server that didn't already had backup.

It's because I get a huge gain from it without additional costs that I use repositories for almost everything - including firewall configurations, todo lists etc.

Have you looked at just downloading and installing CVSNT on your workstation machine?