I store most digitised documents as .PDFs.

Very often you can obtain original .PDFs from companies by downloading from websites, as well as (or instead of) the paper documentation they send you.

For local scanning, I use a HP MFP. If I need to scan individual pages, I can then merge those, if necessary, with a 'merge.pdf' type of software utility.

Store the scanned/downloaded documents in some type of tree-structured directory format. This greatly reduces the time taken to find a specific document.

I keep financial documents separate from other documents. Financial documents are also segregated into separate tax-year 'trees'.

Documents are backed-up month by month, and also daily. The monthly back-ups are stored indefinitely, and separately from the daily back-ups which are deleted in reverse chronological 'exponential' order.

Daily-backups remaining at the moment. Day 0000 was back on 23rd June 2012. Last word is server name. Note how there are more recent backups than earlier backups:

     0000-120623nullius
     1024-150401nullius
     2048-180131centrepoint
     2304-181014centrepoint
     2560-190627centrepoint
     2688-191102centrepoint
     2720-191204centrepoint
     2736-191220centrepoint
     2752-200105centrepoint
     2756-200109centrepoint
     2758-200111centrepoint
     2759-200112centrepoint
     2760-200113centrepoint