Completing completable projects

Something you might not be aware about me is that I am extremely bad at finishing projects. Almost all of my successful projects are structured as anytime projects. They are useful from fairly early on, can be contributed to incrementally, and can be abandoned almost at any point without being left feeling incomplete.

To some degree this is true even with very small notions of “project” - my writing tends to succeed through this method too. I don’t generally write towards a known conclusion, I write until the point where I find a good mic drop moment and hit publish. I’m also very bad at finishing the washing up - it’s not a quantity thing, I can do endless amounts of washing up but somehow the last couple of things don’t get done, or I don’t clean up the sink at the end, or…

This works very well for some things, but very poorly for others. For example I have an ongoing desire to write books, which cannot be structured this way, and every time I find myself trying, it squirms out from under me.

So about four months ago I decided that a good way to work on this would be to create a project which wouldn’t take too long, felt motivating, and had a nice linear path to completion. At a reasonable rate of progression, with a few setbacks, it shouldn’t take longer than two months, and it would be nice to have it completed.

I completed it this past week.¹ The astute observer will notice that four is a larger number than two. Turns out that - surprise! - I am bad at completing projects.

But I did complete it, I am glad to have done so, and I learned a lot in doing it, so I’m here to tell you about that.

Project Outline: Tarot Journalling

The project was built around something I’ve done on and off over the years, which is using Tarot cards as a reflection prompt. I’m a very casual Tarot user (and do not attribute any particular mystical significance to them), but I’ve found that level of casual use very useful.

The basic procedure I use is:

1. Pick a tarot card.

2. Read what Joan Bunning has to say about it (I actually use her Big Book of Tarot, but her website is almost as good)

3. Write something loosely inspired by this (or, if I feel completely uninspired by it, write about why that is).

I’ve done this on and off over the years and it usually goes very well, because writing is so much easier when it’s responding to something rather than something you have to generate from scratch, and it’s a good source of things to respond to that I would not otherwise look at.

This project started when I thought it might be interesting to go through the entire deck, one card at a time, and handwrite an A4 page about each. Initially I planned to do a couple cards per day (my initial version involved setting a 45 minute timer and doing as many cards as I could in that time, which was typically two or three), which put the timeline for the project at about a month, but rapidly decided to scale it down to one card per day, which is where the roughly two months estimate came from.

So, the ideal form of the project was this: My tarot deck lives on my writing desk in two piles, one face up and one face down. Every day, at some point in the morning (at the time I was reserving time before noon for “self work” for want of a better term), I sit down at my desk, draw a card from the face down pile, write an A4 page prompted by it, then transfer the card to the face up pile and put the page in a folder where I was collecting these. The project is done when I’ve completed the last card.

What went wrong

What happened is that, predictably, I fully derailed on the project. I spent about the first week fleshing out the details of how it should work, then it went smoothly for a while, and then it all went horribly wrong.

In mid-June, with a little over ten cards left, I went on holiday and didn’t take my project materials with me, with the intention of resuming when I got back.

Instead, when I got back, a big unrelated event occurred that sortof wrecked me emotionally for a few weeks. During that time, I did try to resume, and almost immediately drew the eight of swords (restriction, confusion, powerlessness) which I absolutely did not want to engage with at that time. I put it back in the deck and drew a new card, but this was probably a mistake as it meant that it was lurking in there waiting for me to engage with it.

As a result, the whole thing acquired an unpleasant miasma of emotional ugh that made me not want to face it. At the same time, the work project that was on came to an end, and I’d been using it to structure my time so it was easy to focus on things like this in the morning. This removed the natural impetus to work on it, making it very easy for me to avoid it, and it became a “ugh I should get back to this, but I don’t wanna” task, in contrast to the easy momentum of the previous version.

Normally that would be the end, and I would abandon the project and feel low-grade bad about it forever. Instead what happened is that after about a month of doing literally almost nothing on it I, very gradually, picked up the remaining pieces, and managed to do in two months what I had originally budgeted less than two weeks for, and now the project is done.

This is, in some ways, actually a better outcome than if the whole thing had gone smoothly. Being able to do something isn’t about being able to do it if everything goes perfectly, it’s about the being able to deal with setbacks, and I learned a lot about doing that in the course of this project.

Which is not to say that I’m suddenly able to complete projects no problem, but I do think I’ve at least made steps towards it.

Recovery from derailing

So how did I fix this?

Well, part of how I fixed it is I wrote an, if I do say so myself, pretty decent article about procrastination, and even though I didn’t really directly apply anything from it, writing it probably helped shift my view of the problem. Derailing is, ultimately just a particularly severe form of procrastination.

In this case, the particular fix was this: If what you’re doing isn’t working, do something different.

You remember what I was saying about how the ideal form of this looked? One card randomly picked, every day, writing desk, done in the morning, full concentration, etc. etc.

Anyway, step one of fixing this was to give myself absolute guilt-free permission to not do any of this.

Here are some things I did in the course of finishing the project:

• Most days I didn’t do anything (clearly, otherwise I would have finished long ago).

• I wrote on days where I was inclined to write, or at least where it felt viable. I have no particular explanation for what made a given day viable, but I paid attention to the feeling of it and took advantage of it when it was present.

• On days that I did write, I wrote as much as I wanted. This was usually still just one page, but sometimes (especially towards the end) it was two or three.

• I very rarely did it at my writing desk.

• I often had a video on while doing it (Emotional regulation strategy, often something I do when there’s a task that I struggle to bring my full attention to for whatever reason²).

• I stopped picking fully at random and picked and chose which card I wanted - at first I picked randomly with the option to try again if it didn’t land for me. Towards the end, I actively sought out particular cards that I felt like doing. At some relatively near the end, this even included actively seeking out the eight of swords (the card that helped derail everything).

The individual details don’t really matter that much in any generalisable sense. The point is that I redefined success around the core concept: One A4 page of journalling, written for each card in the deck. Ideally of good quality, but as long as I earnestly engaged with the topic the quality doesn’t matter, and as long as I achieved this basic standard, the page was a success, and as long as I got all of the cards done, the project was a success.

And it was.

Mode shifts

I think the biggest insight from this is that towards the end of a project there’s inevitably a mode shift that happens at some point, where you stop breezing along doing whatever comes naturally and have to switch to to a different way of engaging in the project, and the strategy for working on it changes.

This was particularly acute and abrupt in this project, which made it easier to notice, but I think it’s a fairly general pattern. I was talking about this on Discord, and Nelson Elhage made the following observation:

I think of the "finishing the project" as a much more deliberate mode of execution; for most of my projects for most of their lifecycle I'm doing some version of following my curiosity or interest, or "taking the next step" (or maybe "a next step" since the project often branches and expands); Finishing A Project is often the point at which I switch to writing down a task list and explicitly triaging and developing a burndown list of things to do (or deliberately not do).

Maybe said another way, Finishing The Project is often where I switch to backwards-chaining from a goal as my primary mode, instead of something more exploratory or driven by intrinsic motivation/interest.

This seems right. The early stages of a project has a huge amount of flexibility and space to manoeuvre, while at the end you need to figure out how to wrap everything up, decide what “done” looks like, and how to achieve that.

I suspect that often when this happens it does look like some prompting event like mine which derails everything. Constrained systems are more fragile, so many things that would not derail the project early will cause problems later.

The rest of the time it just looks like the project has gotten unaccountably harder and stopped being fun, and it peters out and you (I) feel bad about that. This feeling of guilt then creates its own need for a mode shift, because the project now has to have an emotional management component that was previously unneeded, and this means you need to change strategies.

This realisation that projects have mode switches in them that you made need to make at various points is probably somewhat protective against derailing in advance, as it lets you recognise that.

Goals as motivations

I think there are a number of different mode shifts in the course of a project, but the one I’m most interested in is nelhage’s observation that the big one for finishing is switching to backwards chaining from the goal.

I am, honestly, quite bad at goals. Almost everything I do ends up being very… goalless isn’t quite right, but designed with a number of independent benefits, and something that I can just do and reap the benefits of with no clear pass/fail criteria. The idea of doing something and it being for a single well-defined thing is, honestly, quite uncomfortable.

This is not to say that I never do anything motivated by goals, I do, but it tends to be more by way of an action generator. The goal suggests things that you can do that would help for that goal, and I then select from those options things that can be justified as valuable without reference to whether that goal is actually achieved.

This isn’t a bad strategy exactly. It generates a lot of valuable results and flexible tools, but it does run into one crucial problem: It often doesn’t actually achieve the goal, because most of the non-goal value of the action is extracted before the goal is achieved. If you structure your goal achievement strategy as an anytime algorithm, you’ll have captured 80% of the non-goal value with 20% of the work, and then act surprised when the the cost/benefit ratio gets 16 times worse and you suddenly stop wanting to work on the thing.

The solution does suggest itself, which is to treat actually achieving the goal as motivating in its own right. I don’t know about you, but honestly that feels pretty anxiety provoking to me. I don’t exactly know why. I think possibly partly because goals are things you can fail at, and trying and failing seems worse than abandoning the whole thing. Something I need to poke at a bit more.

Finishing articles

As mentioned up above, I typically write in a linear flow, and stop writing when I hit the mic drop moment. This works great, except that it means that most of my articles are completely unedited³, but often it fails catastrophically and results in things lingering in drafts forever because I lost the thread of the writing, couldn’t figure out where they’re going, and as a result didn’t feel able to finish writing them.

This is a perfect example of the sort of mode shift I’m talking about, and one I find starting to happen here: I got most of the above written and then didn’t really know where I was going. This means it’s time to switch from a pure writing mode to a getting the article done mode.

The problem is I don’t have a particularly nice, tidy, conclusion. These all feel like a really dumb set of problems to have (just for me to have you understand. I know plenty of other people with these problems and there’s nothing wrong with that and none of you should feel bad about it. But I’m unique and special and allowed to feel bad about it for… reasons. Yes, reasons. Good ones, I promise), and that makes the whole thing harder to engage with.

But, dumb or not, they do feel like problems that I do now have a little bit more of a concrete handle on, and that can make further progress on by completing more well-defined projects. Recognising that mode shifts are an inevitable part of the life cycle of the project, and giving myself permission to explore different strategies when when old ones stops working, seems like it will give me a bit more breathing room to finish projects. And articles.

1. Depending on your definition of done, I either completed it on Friday, or I complete it when I click publish on this newsletter issue.

2. Including right now while writing this.

3. They’re often edited after I click publish. The astute observer may notice that editing a published article is an anytime project, but editing an article in order to get it publishable is not.