Dynamic Planning

One part of David Allen’s Getting Things Done system I’ve discarded is the idea of sorting next actions by physical context bins, such as phone calls, paper work, computer work, etc. Maybe that makes sense if you travel 200 days a year or work in a high-interruption environment where you can’t concentrate for more than 30 minutes at a time, but given that I work in a home office with virtually all of these contexts within easy reach, I find it worsens my productivity to sort actions by physical context. I get good mileage out of batching errands where I must physically go out, so I do maintain a separate errands list, but otherwise I’ve dumped this part of GTD.

The problem with sorting actions into context bins is that you scramble actions from different projects together. Perhaps you make 5 phone calls in a batch, each of which is associated with a different project. That’s fine if you’re out of the office, and you want to put your cell phone to good use, but what if you’re at your desk in your office? Does it still make sense to batch phone calls just because they all involve physically picking up a phone? If the calls are unrelated, then I’d say probably not.

My preference is to focus on a single project for as long as possible, doing a variety of actions in a row. Once I’ve loaded a project into my brain’s active RAM, I don’t like to unload it. Much efficiency is lost in the process of rebuilding awareness of a project. If I haven’t worked on a project for a while, it can take me anywhere from 15 minutes to several hours to fully reload the project into my brain — this is especially true for technical work or very large and complex projects. So I’d rather work on one project all day long instead of doing a smattering of different actions from ten different projects. I realize that not everyone has the luxury of doing that, but I do.

Taking on too many projects at once and switching between them often during the day will limit the complexity of the projects you can handle. This is fine for simple projects or if you do cog-like work, but it kills productivity on large projects where you need to keep a lot of information in your head at once. Some examples of the latter would be designing a new computer game or a web site, writing a book or screenplay, or doing strategic planning work for a business. In order to work productively on such projects, you can’t keep switching between projects, or the work will take forever. You need to load up a single mental context and stay with it for a long time, preferably days at a time but at least for several hours. Minor interruptions are OK, but you want to keep yourself from having to re-load a whole other mental context. Imagine writing 10 different books at once, working for 30 minutes a day on each one. It would be much more productive to knock off one book at a time.

Sometimes the mental context is a lot more important than the physical one. Sticking with a single project and moving all around your office building to perform the different physical actions may be better than staying at your desk and doing desk work from 5 different projects. Getting up to do something in another room may cost you a couple minutes, but switching projects will often cost you a lot more. When are you working most productively on a project? Definitely not during the first 15-30 minutes.

I think this is one of the hidden causes of procrastination. What happens when we procrastinate? We put a certain project off to the last minute, so we end up having to do the whole thing (or a large part of it) in a single marathon session. Say you put off doing a school paper until the day before it’s due. By procrastinating you ultimately force yourself to do the entire thing in one session. You load the mental context once, do all the next actions in sequence, and then you finish and release the context. This is very efficient in my opinion, a lot better than spreading the work out across several weeks and doing just a little bit each day (and forgetting must of the understanding you gained during the previous week). This is how I did assignments when I was in college, and I managed 31-39 units per semester. If I had a big project, I’d allocate a whole day to it and do it in one session — do the reading, research, writing, editing, etc. If I had to do a book report, I’d read the book and then write the report immediately afterwards. If a teacher allocated a month or two for a big assignment, I’d still try to do it in a single session.

I no longer maintain a separate next actions list, although I used to. Now I keep only a projects list, and I dynamically break it into next actions as needed. For some projects I make detailed plans of all the next actions, but for most projects I break down just enough actions to fill a day or two, and then I do them. Once I’ve completed those actions, I figure out more next actions and then set about doing those. I find this to be a highly productive balance that avoids underplanning on one side and analysis paralysis on the other.

Think of this as dynamic planning. I don’t tend to plan out the details of a project until it’s on my doorstep, and I aim to handle only 1-3 projects at a time. I barrel through, get them done, and then it’s off to the next project. This quote from Tryon Edwards basically sums up my approach:

Where duty is plain, delay is both foolish and hazardous; where it is not, delay may provide both wisdom and safety.

In other words, if you at least know what to do today, then go do it. Go back to planning when you hit the edge of the fog again. Dynamic planning: plan, do, repeat.

If a project is really, really big, then I’ll break it into subprojects, and the subprojects will be scheduled accordingly. Also, since there are always little things to do that don’t fall into any major project (like paying bills and such), I batch those little things up and then dispatch them in a marathon session too. For example, one Saturday I spent the whole day doing 20 unrelated home repairs. And often I’ll write a few days worth of blog posts in one session, scheduling each post to go live on a different day.

The downside to working like this is that once I’ve loaded up a particular mental context, it’s hard to let it go. I become semi-obsessed. My phone will ring, or my wife will walk into the room and talk to me, or my son will be crying in the next room, and I’ll automatically tune them out. It’s as if my brain has allocated all available RAM to the given project, and nothing else will fit. If anything else tries to squeeze in, I’ll chase it away with a growl. On the other hand, if I’m spending a day out with my family, I’m usually fully there with them, not thinking about other projects at all.