084 Goad logo 084 GoadA daily nudge forward
Goals

The Planning Fallacy and How to Estimate Your Goals

I told myself a side project would take six weeks. It took eight months. The code was not the problem and my motivation was not the problem.

A notebook page with a goal timeline sketched out beside a laptop showing a calendar quarter view

I told myself a side project would take six weeks. It took eight months. The code was not the problem and my motivation was not the problem. The six weeks was the problem, and I had picked that number in about fifteen seconds while washing up, which is roughly how most goal deadlines get chosen.

This matters more than it sounds. A goal with a wrong date does not just finish late. It fails in the third week, when reality and the plan separate far enough that you stop looking at the plan, and once you stop looking at the plan you are not working on a goal any more, you are just busy. So the estimate is not admin. It is the part that keeps the thing alive.

The error has a direction

I kept a spreadsheet of every personal goal I gave a date to between January 2025 and early 2026. Thirty one entries. The median overrun was about 1.8 times the original estimate. Four finished on or before the planned date, and in three of those I had padded the number on purpose because a previous version had embarrassed me.

What stands out is not that I was wrong. It is that I was never wrong in the other direction. Nothing took half as long as I thought. That asymmetry is the whole story, and psychologists have had a name for it since 1979: the planning fallacy. People asked to estimate a task imagine the version where nothing interferes, then report that as the expected case rather than the best case.

At work I have some defence against this. Two engineers argue with my number, a sprint board records what actually happened, and a project manager remembers that the last migration slipped. Alone, planning a goal in my own kitchen, there is nobody to argue, so the fifteen second number stands.

Count the weeks you will actually get

The second failure is smaller but easier to fix. We plan against calendar time and then live in available time. A quarter looks like thirteen weeks. It has never once given me thirteen usable weeks.

Here is what a normal thirteen week stretch costs me, measured rather than guessed, from the same spreadsheet.

What eats the quarterTypical costWeeks left
Thirteen week quarter13.0
Travel, weddings, visits home1 week12.0
Illness, household emergencies0.7 weeks11.3
Crunch at the day job1.5 weeks at reduced capacity10.5
Ramp-up on a new kind of task0.8 weeks9.7
Admin the goal creates itself0.9 weeks8.8

Call it nine weeks out of thirteen, or about seventy percent. That ratio has held steady enough across two years that I now use it as a default. If a plan needs more than nine good weeks in a quarter, it is not a plan for a quarter.

The useful side effect is that this converts a vague feeling of lateness into arithmetic. When I am five weeks into a quarter and three weeks of work are done, I am on pace, not behind, and knowing that stops the quiet panic that makes people abandon things.

Where the standard advice fails

The usual instruction is to break the goal into tasks, estimate each one, and add them up. I have run that process dozens of times and it has never produced a number I could trust.

The reason is structural. A bottom-up estimate can only include the tasks you can see from the start, and at the start you can see maybe two thirds of them. The missing third is not glamorous: a dependency that turns out to need a licence, a form that takes eleven days to come back, the week spent redoing something because you learned how it should have been done. In my project logs, work that was not on the original list accounted for somewhere between thirty and forty five percent of total hours. Adding up a list that is thirty percent incomplete gives you an answer that is thirty percent short, and dividing the tasks more finely does not help, because you are subdividing what you already thought of.

What does work is cruder. Find three past efforts of roughly the same shape and look at how long they really took, start to finish, including the dead weeks. Then use the longest one. Not the average, the longest, because the longest is the only one that already contains the surprises you have not thought of yet.

A friend wanted to reach conversational Spanish in four months. Her bottom-up plan said thirty minutes a day for 120 days. Her reference class was three people she knew who had done it, and their real answer was fourteen to twenty two months of inconsistent practice. She reset to eighteen months, kept the thirty minutes, and is still going. The plan that said four months would have ended in March.

Breakdowns are still worth doing. They are for sequencing and for knowing what to start on Monday, which is a different job from duration, and the method in our guide to breaking a big goal into startable steps is the one I use for that part.

A version you can run in twenty minutes

Write the finish line as something observable. Not "get fitter" but "run 10 km without walking". Someone else should be able to look at it and say yes or no. If they cannot, you have no way to tell whether you are late.

Name three reference cases. Other people, earlier projects of your own, published timelines, anything with a real elapsed duration attached. If you cannot find three, the goal is more unusual than you think and deserves more padding, not less.

Take the longest, then convert to available weeks. Multiply by 1.4 if your week is already full, which for most people it is. The nine out of thirteen ratio above is the same move in different clothing.

Set two dates, not one. Engineering teams separate a target date from a commit date, and the gap between them is honest rather than embarrassing. My target is the date I am aiming at, my commit date sits four to six weeks later, and the commit date is the only one I tell anyone else.

Write the number down with the reasoning. One paragraph. In four months you will want to know why you believed six weeks, and future you is a better critic than present you.

When you are already over

Re-baseline once, properly, and then stop. People who redraw the plan every Sunday are not planning, they are negotiating with themselves, and the plan stops carrying any information. Take one evening, apply the reference class method to what is left, write a new commit date, and leave it alone for at least a month.

Then cut scope rather than standards. A thing that is seventy percent as big and actually finished beats a full version that stays at eighty five percent forever. I have shipped exactly one side project out of six, and the one I shipped was the one I cut in half in week nine.

If the goal belongs to a quarter, the thirteen week structure in the 13 week plan gives you a natural checkpoint for that single re-baseline. And if you are carrying four of these at once, the problem is probably load rather than estimation, which is a different conversation covered across our goals section.

FAQ

How much padding should I add to a goal estimate?

Multiply by 1.4 to 1.8 if you are working from your own guess, and by less if you are working from a real comparable project. My own median overrun was 1.8, which is why I distrust anything under 1.4.

Why do bottom-up task estimates come out too low?

Because the task list is incomplete at the start. Unplanned work made up thirty to forty five percent of hours in my project logs, and no amount of splitting the known tasks reveals the unknown ones.

Should I tell people my target date or my commit date?

The commit date. Share the one you can defend, keep the ambitious one private, and you get the motivation without the awkward conversations in month three.

Is it a failure to move a deadline?

Moving it once with a reason attached is maintenance. Moving it every week is how a plan quietly stops meaning anything, and that is the version worth worrying about.

PV
Priya Venkataraman

Priya is a software engineer who plans her day in ninety minute blocks and writes about what happens when the plan meets real meetings.

More posts by Priya

More in Goals

01doorganico 3hot8 4465456 after5internprogram ahandfulofblessings ahmadpc ahomcybersecurity allodialresources annodex antiquesandperiodcollectibles aotc asianworldkashmir beedlectrics bloggertemplatefree bluefeatheragency bodyweightsolution campaign08blog carfixadvisor catstreetcollective cbstvsports cnfanspurchasing committeeof500yearsofdignityandresistance covertopsdjs cs-kaiyuangame cuestacomunicaciontotal cxoutlet dancebetgame deskgearhub destinyvillastay detectmagazine di33 diamondsteelbuildings donmo dyanadropship emarsysfinance emimattl escapechan eskitarih evautoplay everythingweddings-orlando farmzone feedviewerapp foxstudiosscratch gamevui123 gleasongrade greetingsday hacklabasu haskell iw-kaiyun jagoanproperti jilidskripsi joltbusinessfinance juegodepokergratis juegospucca kaeru-festival latestmehndidesigns lestanzediottaviano litemagazines localcommunityhero lustrsazan mariashebets menswheels newwebstechs1 nikkit ootheday oxigenvpn plurfitness prefetial probudgetcarpetcleaning psychicmrsmartin pyaten qvantor rejoignezlogis rexzbot riselayouts ruchin rwxqsgd5 sakaltime shibiaowang shop-bot skrtips spotlesshaven strcoding tetapatrium topselller uneplanteartificielle vipertechcr webtech1223411 wo-kygaming yamaha-mekar yeknava yoybuysheets zenunwa yamaha-mekar jilidskripsi ankertoto ankertoto ankertoto