Next action list categories

Omer Cohen

In the book, David recommends sorting the next actions in the next actions list into categories by context (At computer, office, home etc.). As he says, that is not perfect for anybody - for example, currently, I do 95% of my work at my office. Moreover, I decided to separate my personal life task management into a totally different one.
What kind of next actions categories do you suggest? I thought about a period of time to complete sorting (less than 15 minutes,15 minutes - hour, 1 hour to 3 hours and a day for example) but i am not sure about it?
Will be happy to hear some thoughts.
Thank you very much.


I do to nearly all of my Next Actions from one place: my computer. I sort them by age, that means newest first. Like in a blog for instance, where the newer stuff is on top. I do have one principal NA list for that, one category so to speak.

I usually start to work from the top, because the fast moving stuff tends to be there. Sometimes I deliberately start at the bottom of the list, where the slower stuff accumulates automatically.


Then you could divide the list in more than one category: @worked-on-this-week, @worked-on-last-week, @worked-on-prior-to-last-week, @never-worked-on.

So, you created a new project, the Next Actions for that project go to @never-worked-on. When you are working on a project and are now finished working on that project for now, the new Next Actions you now create goes into @worked-on-this-week.

During the Weekly Review, you pull all Actions from @worked-on-last-week over to @worked-on-prior-to-last-week. Then you pull those from @worked-on-this-week over to @worked-on-last-week.

Note, this is just one of many possible variations of this.

Omer Cohen

Then you could divide the list in more than one category: @worked-on-this-week, @worked-on-last-week, @worked-on-prior-to-last-week, @never-worked-on.

So, you created a new project, the Next Actions for that project go to @never-worked-on. When you are working on a project and are now finished working on that project for now, the new Next Actions you now create goes into @worked-on-this-week.

During the Weekly Review, you pull all Actions from @worked-on-last-week over to @worked-on-prior-to-last-week. Then you pull those from @worked-on-this-week over to @worked-on-last-week.

Note, this is just one of many possible variations of this.
That's very interesting. What do you think about my idea of period of time to complete the task?


The main problem with that is that you burden yourself with the task to estimate a duration time for every NA and this creates friction. Suddenly you can't just add an NA to your system, you also have to estimate and write that down.

Additionally, often times our estimation will be wrong, which is not a problem originally. But now it becomes a problem. If you estimate wrongly you are attacking the integrity of your system. You will undermine your trust in your system, if you repeatedly make the experience of wrong estimates. You'll stop believing them.

Also, there is a trap: if you estimate a lot of effort for a given NA, this could act as a demotivating force, which leads to procrastination.

What I can see working would be a list of NAs that will *absolutely* take a bigger block of time, say, 2 hours or so, and then block out time on your calendar to tackle one of those.


The obvious criterion is by urgency. In the standard GTD set-up you already have that in a way. All NAs on the @context lists are ASAP (as soon as possible) and everything else is either more precisely sorted in the calendar and tickler file. On the other end of the spectrum are @WaitingFors and Someday/Maybes, things whose urgency is comparatively low.

Now, does it make sense to sort your ASAPs by further degrees of urgency? If you are doing it right, the more urgent stuff already is on the calendar and the tickler file functionality. If it is not to be tackled this week, why even have a NA? Why not put it on Sd/Mb?

As far as I can see - and of course I could be easily wrong - the standard GTD set-up provides division by urgency in a very good granularity:
  1. calendar / tickler file
  2. ASAP, @context lists
  3. @WaitingFor
  4. Someday/Maybe.
I think this is pretty good.


Others I've seen sort by Area Of Focus, which I like in a way, because it allows for making sure you don't neglect any area of your life. Years ago there was an app, Life Balance, which would do this for you automatically. You could tell the app how much you wanted to work on each area and then it would sort your lists according to that.

I never tried this out, so I have no idea how well this would work in practice.


What kind of next actions categories do you suggest?
I've tried both putting things in by Area of Focus and by time to complete. In both cases it became much more effort to enter in new stuff and I fell behind.

For me I separate actions into contexts based on what application I need to use (LibreOffice, Scrivener, LightRoom etc.), which computer I am most efficient at doing that action on (Main mac for e-mail, reading on my iPad, Android development on the laptop), which barn or building I need to be in (Shop, Hay barn, Red Barn, Little House), which town or city where the errand can be completed, (Local town vs big city 75 miles away) whether I can do it at any time or only business hours (Phone Business and Phone), whether I need assistance to do the action, (Inside with Help, Outside with Help) and so on.

I also create, use and delete contexts as needed.

I use my weekly and in depth quarterly reviews to make sure I am getting projects done in all my areas of focus.


I think of these things as potentially orthogonal labels or tags - but then, because that gives us combinatoric explosion, we extract certain buckets that contain reasonable numbers of reasonably related items.

E.g. I need to strictly separate personal stuff from work stuff.
Category=Ownership, Values=Personal,Employer

Sometimes more than just Personal/Employer, because I may still be obliged to review patent applications for Past-Employer(s). But usually there aren't very many of these, so I just fold them into the Personal bin.

E.g. some of my work requires that I be Online, connected to my company's network by VPN. Some company work can be done offline, e.g. preparing PPT slides or writing email (at least up to the point that I need to search the company intranet). Some of my personal stuff can be done offline, some requires onlije access.

E.g. some stuff can be done on my SmartPhone, some requires me to sit at a PC.

This might result in tuples like



But not all combinations are common. E.g. while there could be categories for (Work,SmartPhone,Online,VPN), my SmartPhone screen is too small for me to read most technical email.

And some combinations are currently impossible. E.g. I currently do not allow VPN into my personal home network.

Even this simple set of categories

would require 2x2x3 = 12 "completely expanded" categories.

Then add some more, like the GTD

Let alone

Full combinatoric expansion gets us 2x2x3x4x2x3x...

I can't keep track of so many. Not in separate lists.

A computer could - but again, I don't want to see them.

So, the way I think of it, is that we just highlight, separate out, certain bins that are particular important right now.

And push everything else into bins with less fine grain distinctions.

In an ideal world, with the SW tools I dream of, we might be able to "re-sort", change the bins - and use tags already assigned to create a different set of lists.


Sounds like you want multiple labels instead of contexts, and then get dynamically generated next action lists on the fly by selecting the appropriate labels. That shouldn't be too hard for existing software. But I'm not 100% sure that this is actually useful.

The purpose of a context is to help you narrow down the huge list of next actions to a list of tasks that is of reasonable size and that you can perform with the tools you have and the place you are at.

However, that would sometimes lead to an unreasonable amount of fine tuning. In my opinion, if your action lists are reasonably small, it's not the end of the world if the odd task makes it on that list that isn't doable in the broadest sense of the context.

Stephen Dewitt

I also work the vast majority of the time with everything I need to hand (laptop / smartphone) so most of David's contexts aren't that useful - however he does also mention energy level in the book which has become my main categoriser (high, medium, low, snoozefest). I have a lot of energy in the morning, less in the afternoon, sometimes very little at all - so I check my high energy list for big things like write ups in the morning, then do either admin or spreadsheet type stuff in the afternoon.

In terms of energy required I take into account how much I like the task also (so I think it is a product of [Willpower required]x[Cognitive difficulty of task]) - if I am writing a paper I am really enjoying, even if it's challenging, I could maybe do that in the afternoon, as my enjoyment will overcome my tiredness, but sometimes a relatively simple task like an email I just really don't want to write / send could be a 'high' energy task if I have a strong gut reaction against doing it - I'll have enough energy in the morning to overcome that aversion.

Also I have a definition of cognitive difficulty I really like - snoozefest = I could do this task while listening to an audibook / podcast, low = I could do it with music with lyrics or casual radio, Medium = music without lyrics e.g. classical, High = background noise only (/ 'background sounds' app) -- some grey areas of course.
Last edited:


IMHO duration of next action is not a context. Next action shall be described by four elements: context, duration, energy, and priority.
List of contexts depends on one's activity. Even if your work happens mostly in one place you can still be using context. I think that behind context conception is one idea - changing context costs some mental / physical energy. So constructing contexts list you can realy among others on that citerium.


It’s all well and good to have ideas for contexts based on theories and personal preferences, but the central question is what works best, in the sense of better outcomes at an appropriate pace. Because we are influenced by factors at a less-than-fully-conscious level, it is necessary to test changes to our systems in the real world. This also requires the honesty and self-awareness to admit when something isn’t working out perfectly. For example, I found that an @email context just slowed me down, so I folded those next actions back into @computer. It’s not a question of wrong or right, but what works best for you.
Last edited: