The email you swore would take five minutes
You know the moment. You sit down with a short list and a clear head. The first task — answer an email, draft a paragraph, fix one small thing — looks like it should take five minutes, maybe ten. An hour later you are still inside it, slightly bewildered, wondering where the time went. Then you do the same thing tomorrow. And the day after that.
This isn't a character flaw, and it isn't laziness. It's one of the most reliably documented biases in all of psychology, and it has a name: the planning fallacy. Once you understand the machinery behind it, you stop being surprised by your own calendar — and you start planning in a way that actually survives contact with reality.
What the research actually found
The term comes from Daniel Kahneman and Amos Tversky, who noticed something strange in the 1970s: people's predictions about how long their own projects would take were systematically too optimistic, even when those same people had been burned by the exact same kind of project before. Knowing you ran late last time did almost nothing to correct this time's estimate.
Later work by Roger Buehler, Dale Griffin, and Michael Ross put numbers on it. In one well-known study, they asked students to predict when they would finish their honors theses — both a realistic estimate and a worst-case, everything-goes-wrong estimate. On average, students finished later than even their worst-case predictions. The pessimistic guess was still too optimistic. Read that twice: people's worst-case scenarios turned out to be more accurate descriptions of the typical case.
The effect holds across kitchen renovations, tax returns, software launches, and the laundry. It is remarkably stubborn.
Why your brain does this
The core reason is what researchers call the difference between the inside view and the outside view.
When you estimate how long something will take, you naturally take the inside view. You imagine the task unfolding step by step: I'll open the document, write the intro, fill in the middle, tidy it up. You build a smooth little movie in your head, and you time the movie. The trouble is that the movie only contains the steps you can picture — and it never includes the interruptions, the misremembered detail you have to go look up, the file that won't open, the phone call, the moment you realize the whole second half needs rewriting. Those things are real and frequent, but they're unspecific. You can't picture which one will happen today, so you leave them all out. The result is an estimate built from a best-case story.
The outside view ignores the story entirely. It asks a colder question: how long have tasks like this actually taken me in the past? That number quietly contains every interruption and detour, because they happened, even if you can't predict which ones recur. This is why the outside view is so much more accurate — and so much less appealing. It feels generic. It refuses to honor how clear and simple this particular task looks from the inside.
There's a second culprit, too. We tend to remember the focused minutes of past work and forget the dead time around them — the settling-in, the re-reading, the small recoveries between bursts of attention. So even our memory of how long things took is shaded optimistic, which feeds the next estimate.
The trap of breaking things down — and the way it helps
Here's a subtle wrinkle. Research on what's called the segmentation or unpacking effect shows that when you break a task into its component parts and estimate each one, your total estimate goes up — and usually gets closer to the truth. Listing the sub-steps forces you to notice work that the single smooth movie glossed over.
But unpacking has a limit. It surfaces the steps you know about. It still can't surface the unknown unknowns — the surprises that, in aggregate, are almost guaranteed to show up even though no single one is predictable. So breaking a task down helps, but it tends to fix the visible underestimate, not the invisible one.
The most robust correction in the literature is reference class forecasting, an idea championed by the planning researcher Bent Flyvbjerg and rooted directly in Kahneman and Tversky's outside view. Instead of asking "how long will this take," you ask "how long did the last several things like this take, and what should I budget on that basis?" You forecast from your own track record, not from your hopes.
How to actually plan against it
You can't delete the planning fallacy. It's wired in, and it returns even when you know about it. What you can do is build a few habits that quietly absorb it.
Estimate, then look back, not forward. Before guessing how long a task will take, ask when you last did something similar and how long it really ran — including the messy parts. Let that memory, not the clean mental movie, set your number. This single move does most of the work.
Plan in real units, not vibes. "This morning" and "a quick thing" are not durations; they're moods. Convert your work into countable blocks of time. When a task is a fixed-length unit — a focused stretch of, say, twenty-five minutes — your plan becomes a concrete claim about reality. "This will take three blocks" can be checked against what actually happens. "This is quick" can't.
Track the gap. Note your estimate, then note the truth. After a week or two you'll have a personal correction factor — maybe things reliably take you 1.5 times your first guess. Apply it on purpose. You are now reference-class forecasting on yourself, and it is humbling and useful in equal measure.
Budget for the unspecific. Since you can't name tomorrow's interruption, name the category. Leave deliberate slack in the day that isn't assigned to any task — not as a failure of planning but as the most realistic part of it. The day that's planned to the minute is the day the first surprise breaks.
Watch the focused time, not the wall clock. Part of why estimates drift is that we conflate "time spent on the task" with "time the task was open in front of me." A task left open for two hours might contain forty minutes of real attention. If you measure the focused minutes directly, your future estimates stop being polluted by the dead time you'd otherwise forget — or, worse, forget to forget.
The quiet relief of planning honestly
There's something freeing about all this. The reason your days feel like they're constantly slipping isn't that you're bad at your work. It's that you've been negotiating with a story your mind tells — the clean, frictionless version of the task — and reality keeps sending the bill. When you plan from your actual track record instead of from the story, the gap between intention and outcome narrows, and the low background hum of running-behind starts to fade. You stop being disappointed by a process that was working exactly as the science predicts.
This is the thinking behind Tally, which is built to let you plan in real units rather than hope. By pairing a Pomodoro focus timer with your habits, it turns vague intentions into countable blocks — so "this'll be quick" becomes "this is two focused sessions," and the time you actually spend gets measured instead of imagined. Over a few weeks, that record becomes your own reference class: a quiet, honest answer to the question of how long things really take. If you're tired of being ambushed by your own calendar, you can start planning against the planning fallacy at tally.lumenlabs.works.