Posted On: 2023-01-02
I have something of a tradition where I begin every year with a blog post about my plans for that year. Yet I've never sat down and scrutinized whether or not that approach is helpful. Thus, today's post is something of a divergence from that tradition: rather than a retrospective and planning session for the project itself, I will open the year with a retrospective for this tradition of annual planning.
Beginning the year with a planning session serves two unrelated purposes:
My first annual plan was for 2019, and it was a bit of an outlier as far as planning sessions go. I was still at the very beginning of this long-term project, and the plan was at once both overly ambitious and rather vague. While I did complete most of the near-term goals (finding the UI framework was the only item that was explicitly cut), the wording is so vague that simply reading them now, it's unclear if they are actually complete (more on that later). The longer-term plans are even worse in this regard: most of them have no defined end-state, so I doubt I will ever be able to say they're done - even once I've finished the project.
The first year was a bit of an anomaly in my pattern since I did a new planning session in March. Reviewing that post, it's clear that the near-term goals defined at the start of the year were useful for that retrospective effort, and that I was satisfied that the near-term plan was successfully completed (minus the UI thing). The plan for March, however, was even worse defined than the plan from the start of 2019 - not least because it was not timed well with respect to actually completed work (I was in the middle of developing a prototype to gather data about what I should focus on next.)
While the plan for 2020 doesn't make any explicit reference to the March planning session, the content of the retrospective focused primarily on events that happened after March, suggesting that planning in March was at least somewhat useful (even if I did later overlook it during multi-year retrospectives.) On the topic of how the 2020 plan itself worked, spreading the planning out over time (and including a second post) was extremely helpful for pinning down specifics. Interestingly, I needed to revisit the plan twice during that year, indicating that, despite the high level of (useful) detail, I still managed to miss critically important elements.
Surprisingly, at the start of 2021, I didn't take any time to do a retrospective, and simply started in on my next plan. While that could be interpreted to indicate that the 2020 plan wasn't useful as a retrospective tool, I think it's more an indication of the tricky spot I found myself in at the end of 2020. The actual plan for 2021 is the most vague of all - perhaps it would be more accurate to simply say that 2021 simply didn't have a (publicly visible) plan. Instead, I had spent my planning time working through a major issue with my story (the "wayward goal" mentioned in the 2021 plan), and creating a concrete plan for the story itself, rather than for the execution of the project. Yet, looking at a retrospective from later that year, writing made up a fairly small amount of the actual work for that year - what I'd planned and what I accomplished were largely unrelated, even if both were individually useful for the project.
2022 followed a similar pattern to 2021, unfortunately. Rather than a lot of small goals (like 2019 and 2020), I chose to pick one really big goal (autonomous agent movement) and dedicate all my time to that. Also unlike 2020, I didn't take the time to (publicly) break that larger goal down into smaller pieces - instead that happened privately over the course of the year, each time in response to discovering new twists or issues with how I was trying to achieve my goal. Looking back on the year now (when I'd normally do a 2022 retrospective), it's pretty clear that this lack of planning was a significant (but not the only) factor in the year feeling profoundly unproductive.
Looking over my annual planning process since the start of the project, it's clear to me that breaking down larger plans into more detailed tasks is essential for productivity, even though it both takes more time and doesn't catch everything. Documenting plans consistently in a single place makes it easier to perform long-term retrospectives (ie. annual or multi-year), but it doesn't appear to have much impact on my ability to measure progress within the year itself. What's more, the most useful plan appears to have been the one that I completely rewrote: at the start of 2020 I intended to write a single post with my plan embedded within it, but by the end of my writing session I'd committed to spreading out the planning over the course of the week - effectively sacrificing meeting an (arbitrary) deadline for making a higher quality plan. In light of this, I intend to do the same this year: there's a lot of work ahead, both short and long term. Neither rushing a plan nor proceeding without one will get me nearly as good long-term results as spending time really digging into it.
Looking at the plans from the start of each year, it's actually quite unclear how much progress was made over the course of the previous year. Retrospectives that are included before each plan generally focus on changes since the previous retrospective, rather than the year as a whole, yet they often omit links to (or even mentions of) that previous retrospective. As a result, I can't rely solely on these planning posts to account for all my progress over the course of the year - instead, I need to comb through my previous posts, searching for retrospectives that are sprinkled throughout in order to get a complete picture of how the year unfolded.
Based on those patterns, I think it's quite clear what I need to do - both for this year's plan and for future ones. Firstly, I need to sit down and inspect, in detail, what work I intend to do. Even when they're wrong, detailed plans are clearly more useful than vague ones. Secondly, I need to embrace the fact that I often look back at my progress at various points throughout the year. Sometimes those occasions are places where I pivot or take on more work (and thus would do well to have more detailed plans), but they just as often can simply be where one task ends and another begins, and the mere transition is interesting enough to pause and reflect. In light of that, I really shouldn't treat these annual plans as a reflection of the work for any given year: they are simply an arbitrary point in time, and there are many other equally (if not more) important points for understanding my progress on my project.
Going forward, I hope that this knowledge will help me to plan better, execute better, and even reflect better on the progress that I make. This is a year of many changes for my project, and I hope that starting it off by examining this (seemingly innocuous) tradition will set the tone for the rest of the year: digging deeper into processes and habits that have accumulated over the years, and finding ways to make them more valuable to the project as a whole.