I don't have "hard rules" for now, I find that depending on the project or area of responsibility, and also the stage a project is at, it will vary. For a newer project, I will spend more time in planning mode, at least 1-2 hours a week, usually at the end or beginning (to wrap up my week or be ready for the week to come). If nothing changed much and the project is moving forward normally, it might be just 5-15 minutes to quickly check that everything is on track, if I have to change some action from "waiting for" to a "follow up" context, etc.
Regarding emails, I try to go through all my emails every day (morning or evening), and try to set a goal to read every email only once, then turn it into an action or a project and file it right away. Doing this removes for me a lot of overhead and also allows me to focus by removing the "unknown" factor of unread emails or emails not completely processed. Similarly, new ideas will go to OF inbox if I think it will be benefical to review it at the end of the day / beginning of next day (something unusual or unexpected for example, that I feel might need a bit more thinking, but without interrupting what I do at the moment).
To work on projects, I find it useful when I have a lot of projects going on to decide to work on only 1-2 projects on some days (knowing that there will be emails and other small tasks for the other projects), ideallly blocking 2-3 hours for each project as if it was a meeting or something like that. I think it definitely help to do that to "dive" in a project. But that is to work or think about a problem, rather than doing the "project review" in which I will not work on project tasks or solve project problems, but rather plan the project and think about project organization (I hope that makes sense!). For example, in the project review, I will think in terms of "this block must be finished before this block", "who can help for this problem, ok I will contact him next week", risks, etc. When I'm working, if something like that comes up, I will add an action to think about this when I'm in review mode; otherwise, I complete actions. So during the review I think of all the actions, their sequence, risk, things to do; and when I work, I just "do".
One way that works well as far as contexts go for my job (computer, hardware/software development), is to use contexts such as:
- VPN (things I must check on client's servers, support, etc.);
- analysis / troubleshooting (more open-ended);
- persons (manager, people I assign work to, etc.)
Then, if I connect to the VPN, I will do all the VPN tasks, but nothing "beyond" what requires the VPN. For example, if I need to check some data in order to perform some other action, I will just check the data, and then move to the next task for this project, which needs the VPN. That's what I mean by "round-robin", it's not so much that it is random, but that I go through potentially a lot of different "action paths" for different goals, but I do only a little part then move to the next. Most of the time, because "connecting to the VPN" represents a high "cost" (when I am connected I cannot run certain software, sometimes the connection itself requires that I call someone to get the authorization, etc.), by connecting once and moving forward on many fronts at once.
And similarly for other activities, if I'm all setup to code, compile, my logs are readily available, I might as well do that for a few different tasks/projects.
So as you can see, the emails, new ideas, etc. that come get quick pushed into the system, and at the end of the day/week (or beginning if you're a morning person), I do the review process. This cuts on the cost of "context switching" (in the "computer" sense, of getting a project in your "RAM").
Some things that I'm still incorporating and I find help a lot is during the review spending time to really break down things into "atomic" actions, and order them properly. The more I do this well, the more the "work part" becomes just going through a list, with a good focus. By selecting a few projects for the day / week, I make sure those move forward. By pushing new ideas in OF, that I know I will review within a reasonable delay, I don't break my focus and don't affect my productivity.
I hope this helps at least a little bit! :)