What we learned this year
An honest inventory from twelve months of work: the lessons, stated plainly, including the ones we had to learn twice.
Every December we sit down and ask what the year actually taught us, which is a different question from what we’d like it to have taught us. The first list is shorter and less flattering. It’s also the only one worth writing down.
So here’s ours, stated plainly, no case studies attached. Some of these we learned this year. Others we relearned, which is its own small embarrassment, and probably the more honest category.
- Most emergencies weren’t. The genuinely urgent things announced themselves clearly; everything else was somebody’s anxiety wearing a deadline. We got better at telling them apart, mostly by waiting an hour before reacting.
- Saying no early is a kindness. Every slow no we delivered this year cost more than a fast one would have. Not one fast no damaged a relationship.
- The phone remains undefeated. Threads that ran for days collapsed into ten-minute calls, over and over, and we kept being surprised by it, which suggests we haven’t fully learned this one yet.
- Scope doesn’t creep. It walks in through the front door while everyone watches, one reasonable-sounding request at a time. The defense isn’t vigilance; it’s a written definition of done that somebody actually rereads.
- Rest is part of the work. The best thinking this year reliably happened after time away from the problem, and we still schedule as if it happens at desks.
- Momentum beats mood. On the days nobody felt inspired, the small dull next step got things moving anyway. Waiting to feel ready turned out to be the most expensive habit in the building.
- The work tells you when it’s done. Ship-it feelings and done-ness are different signals, and confusing them cost us in both directions this year: things held a week too long, things released a week too early.
- Write things down. The decision logs we kept sporadically turned out to be worth more than the strategy documents we polished. Memory is a bad archivist and a worse witness.
The pattern underneath
Reading that list back, a pattern emerges that we didn’t plant on purpose. Almost nothing we learned this year was about craft. The design got better and the code got better, but the lessons that mattered were all about judgment, timing, and the discipline of not doing things. Fourteen years in, that tracks. Craft compounds quietly on its own schedule.
Judgment has to be paid for, usually at retail.
What we didn’t learn
What we didn’t learn deserves a line too. We still haven’t solved estimation, and at this point we suspect nobody has; they’ve just stopped admitting it in public. Internal projects continue to outnumber internal finishes by a ratio we decline to publish. And meetings, despite a full year of pruning, grow back.
The method
A note on method, since a few people have asked how we run this exercise. Everyone writes their list alone first, before any group discussion, because the fastest way to lose the honest version is to let the loudest version go first. Then we read them out and keep whatever shows up twice. Independent agreement is the closest thing an agency has to peer review.
None of this will appear in a capabilities deck, which is partly why we’re putting it here. An agency that only publishes its wins is running a highlight reel, and highlight reels are for recruiting, not for thinking. The journal exists for the other thing: the honest inventory, taken annually, in front of whoever cares to read it. It also keeps us honest next December, when the temptation will be to remember this year as tidier than it was.
As for 2025: no resolutions. Resolutions are deadlines borrowed from the calendar, and we wrote a whole piece about those a couple of weeks ago. The habits that survived this year on their own merits get to stay. Anything that needed constant defending can make its case again in January, to a better-rested jury.
See you next year. Same voice, fewer meetings.
Probably.
