I opened ClickUp on a Tuesday night and “what I need to do tomorrow” had three items on it. I ran a quick reality-check before going to bed. Two of them were already done. One of them wasn’t mine. The lesson — the task tracker is the last thing to know — has rearranged how I plan my week.
The list that didn’t survive contact with reality
I’d had a productive day. 111 commits across nine repositories. The kind of day that makes you a bit smug at dinner and a bit guilty at midnight because if you can do that much in a day, what have you been doing the rest of the week. I opened ClickUp to figure out what to attack on Wednesday.
The top of the list:
Top 3 for tomorrow:
1. WhatsApp Jeremy → ask him to introduce Brahim
2. Send Tobias the TG-website redesign status update
3. Finish the 3-tier AwardVantage pricing docThree items. All flagged “high.” All open in ClickUp.
I sat with them for thirty seconds and ran a small audit out of habit. The audit went, item by item:
- WhatsApp Jeremy — I did this in March. Two months ago. The message is still on my phone.
- Tobias TG status — I sent that last month. The git history on the project repo proves it.
- AwardVantage pricing — that’s not even my task. That’s the other person’s task. The lead it depends on is cold until winter.
Three tasks. Zero of them were “what I need to do tomorrow.” All three of them were ClickUp telling me a story that had no relationship to what had actually happened. My tracker was three months behind reality and presenting that lag as my Wednesday morning agenda.
If I hadn’t paused to check, I’d have woken up and done one of two things: re-done a thing I’d already done, or — more likely — felt vaguely guilty about three open tasks and avoided them by doing something easier. Both outcomes are losses. Both outcomes happen because I trusted a system that was lying.
Why this happens
It happens because closing tasks is a separate motion from doing them. Doing the thing is rewarding. Closing the ticket about doing the thing is administrative. The first finishes; the second slides; and the gap between what I’ve done and what ClickUp thinks I’ve done opens silently, every day, forever.
It also happens because tasks have a half-life and ClickUp doesn’t know it. “Send Tobias the TG redesign status” was a meaningful task in March. By May it’s noise. Nobody is waiting for it; the situation it described has changed; the action it asked for is irrelevant. ClickUp can’t tell. The task sits there with the same urgency it had when I created it, because the urgency I attached was a snapshot of how I felt at task-creation time, not a live signal.
The protocol was sound. My discipline was not.
That’s the version I told myself for six months. It’s also wrong. The protocol isn’t sound — it’s a write-once-read-many design that requires the writer (me) to make manual updates whenever reality changes, and the writer (me) is unreliable about that. The discipline is fine. The tool is asking for something humans don’t reliably do.
The four signals I trust more
Once I stopped trusting ClickUp as a “what’s pending” oracle, I started looking at what I did trust. Four things, ranked by how rarely they lie:
| Signal | What it tells you | How often it’s wrong |
|---|---|---|
| Git commits (last 7 days) | What I’m actually building right now | Almost never |
| Calendar (next 7 days) | What I’ve committed to attend or deliver | Rarely |
| OneDrive / file timestamps | What documents I’ve touched | Rarely |
| My own narration (daily note, voice memo) | What I think is happening | Often |
| ClickUp open tasks | What I promised in the past, regardless of present | Constantly |
Git is the one that surprised me. What I commit is what I’m doing. No room for “but I meant to.” No backlog of stale aspiration. The commit log is the truest record of where my attention has actually been, ordered by when it was there. If the commit log says zero commits to project X in two weeks, project X is not active, regardless of what ClickUp says about its tasks.
The “my own narration” line earns its dishonourable mention. I do daily notes; the daily notes are also a lagging indicator, just by a few days instead of a few months. When I run a triage, I trust git and the calendar before I trust my own writing.
The reframe: ClickUp is for promises, not pending work
I haven’t deleted ClickUp. It has a real job. The job is capturing things I’ve promised in the past so they don’t fall out of my head. That’s a valuable job — without it, I’d forget the small commitments and break trust with people who count on me.
But it’s not “what I should do tomorrow.” It’s a promises ledger, and a promises ledger has different rules:
- Heavy gardening, every Friday. Every Friday, open ClickUp, close anything that’s done, reassign anything that isn’t mine, defer anything that’s blocked. Treat stale tasks as garbage to take out, not as guilt to carry.
- Don’t plan from it. Plan from git commits and the calendar. Look at ClickUp to remember what I promised, but the actual “what’s next” decision comes from elsewhere.
- Allow it to be empty. A clean ClickUp means I’m caught up on promises, not that I’m not working. The promises ledger emptying is a good thing.
The reframe took thirty seconds to understand and several months to actually do, because the feeling of an open task is “I owe someone this” and closing one without doing it feels like cheating. It isn’t cheating. It’s accepting that you did the thing in March, you’ve done many other things since, and the open status is a clerical error.
What I now do on a Tuesday night
The triage I run before going to bed, in order:
- Open
git log --since='1 week ago' --all --author=meacross my project folders. That’s the actual “what have I been working on” report. Five seconds. - Look at tomorrow’s calendar. Anything I’ve committed to deliver?
- Look at the bottom of ClickUp — tasks I haven’t touched in 60+ days. Close or defer. (Almost always close.)
- Then, and only then, look at the top of ClickUp — and read it as “things I promised in the past that might or might not still be live,” not as a Wednesday agenda.
The Wednesday agenda comes from steps 1 and 2, with step 4 as a sanity check. It almost never produces the same list I would have produced from step 4 alone.
The reframe in one line
The task tracker is the slowest part of your system. It tells you what you committed to, weeks or months ago, with the weight you assigned at the time. It does not tell you what’s happening now — git does, the calendar does, the deliverables on your hard drive do. Plan from those. Use the task tracker for what it’s actually good at, which is remembering things, not deciding things.
If you’ve ever opened your task tracker on a Sunday night and felt the dread of three months of accumulated debt — that’s the lagging indicator talking. The debt is mostly imaginary. The work is mostly already done. The tool is the slow one.
Adjacent: I retired my n8n server. Here’s what replaced it. (forthcoming).