Ask a small business owner what happens if they are unable to work for two weeks and the answer is generally a short laugh. Everything stops. Customers are not contacted because nobody else knows who is waiting. Invoices do not go out because nobody else has ever sent one. Suppliers are not paid because only one person has the bank access. Work in progress halts because only one person knows what stage it is at and what was promised. The business does not fail in a fortnight, but it takes a month to recover from one, and the recovery falls on the same person who was ill.
The reason this is so rarely addressed is not that it is hard. It is that addressing it means sitting down and admitting how much depends on one person, and that is an uncomfortable afternoon. It also feels like preparing for something unlikely. It is not especially unlikely - a broken bone, a hospital stay, a family emergency, a serious bout of flu are all ordinary events - and the discomfort is precisely what stops a task that would take a couple of hours from being started. This is worth naming, because the obstacle is not a lack of time and treating it as one leads nowhere.
The practical version is not a continuity plan in any formal sense and should not be made into one, because a formal document is the thing that will never get written. It is a list, kept somewhere a trusted person can reach, that answers a small number of questions. Who are the customers currently waiting on something, and what were they promised. What must be paid in the next month, and from where. Which accounts and systems are needed to keep things running, and how does somebody get into them - by a route that exists rather than a password in a drawer. Who should be told, and what should they be told. Who can make a decision if one has to be made.
The access question is the one that most often turns out to be unsolved, and it is worth thinking about now rather than improvising later. Sharing passwords casually is not the answer, and neither is an envelope nobody has looked at in three years. Most password managers provide a proper mechanism for granting access to a named person in defined circumstances, and a bank can add a second authorised signatory, which is a form and an appointment rather than a project. Both are boring and neither takes long. What they produce is not protection against the event, which is not available, but a fortnight that costs the business a fortnight rather than a quarter - and, for the person who is ill, the ability to actually be ill.