There are three ways that IOG integrates: either with the charge point, or with the car, or both.
If like me, it’s with the charge point (Ohme Home Pro) but not the car (e208), then a starting point with any visitor is just to do the pro rata kWh battery arithmetic: e.g. If mine is 50kWh and the visitor’s is 80kWh, then I ‘pretend’ the car is still an e208 but ask for 1.6x what the visitor needs to add. So if they need to add 20% then I ask to add 32%. I then set the ‘Ready by’ time (which can be anything, including outside the fixed 23:30-05:30) and plug it in.
With IOG+Ohme you always have this option, even if the car has integration with Octopus. And because some models seem to have flaky integration (not least because manufacturers are prone to changing their software, so it breaks), some owners opt to take this ‘add x%’ approach.
The advantage of integration between Octopus and the EV is that, as you point out, you can just set a target charge% and that’s it. Octopus is able to ask the car or the car’s app what its state of charge (SoC) is, so as the charge session progresses it stops as specified. But this requires regular polling of the EV and in some models, this happens so often that it exhausts the 12v battery. So again, these owners opt to use the ‘add x%’ approach even though in theory, they should be able to use the ‘charge to y%’ approach.
If you do use the integration with your main car and Octopus, then this is great until you need to switch to another EV. In which case you need to tell the system that it’s a different car (I think, but happy to be corrected). And then afterwards have to reset everything back. This is complicated further if the other car isn’t integrated with Octopus, as you have to switch to using the ‘add x%’ method.
As pointed out, the trouble with establishing a reliable charging regime is that there are a lot of different combinations. And the difficulty varies by EV model, charge point model and tariff. And there are often pros and cons to each. In reality, outside of software bugs, it’s not that difficult to arrive at a workable solution. But it can look a lot more complex than it really is.