ADHD Time Estimation: How to Know How Long a Task Will Actually Take

"It'll take twenty minutes." You said it with total confidence. Two hours later you're still in it, three other things you promised have slipped, and you're the unreliable one again — even though you worked the entire time.

This is the part of time blindness that does the most damage to your actual life. Losing track of time during a task is annoying. Misjudging how long things take before you start is what wrecks schedules, blows deadlines, and makes you look flaky to people who don't know you were heads-down the whole time.

The standard advice is "add a buffer" or "multiply by 1.5." That's not wrong, it's just untethered — 1.5x of a guess is still a guess. This guide is the version that works: stop trying to estimate better, and start measuring instead, until you have a personal multiplier built from your own data.

Why ADHD Time Estimation Fails

It's worth understanding the mechanism, because the fix follows directly from it.

You have no reference data. Estimation is fundamentally a memory task — you predict how long something will take by recalling how long it took last time. But if you don't reliably perceive duration while it's happening, you never encode an accurate memory of it. So you're not estimating from experience; you're estimating from a vague feeling of how long the task should take. Garbage in, garbage out. ADDitude's piece on why time estimation stumps ADHD minds describes this as time blindness warping the internal ruler you'd use to measure with.

The planning fallacy hits you harder. The planning fallacy — the well-documented human tendency to underestimate how long our own tasks will take, described by Kahneman and Tversky — applies to everyone. Neurotypical people are optimistic about time too. But they have a roughly functioning internal clock correcting them, and you don't, so the same bias lands on a much shakier baseline.

You estimate the work, not the wrapper. When you picture "reply to that email," you picture the ideal version: you, already focused, typing the reply. You don't picture finding the thread, re-reading the context, looking up the attachment, getting derailed by a second email, or the four minutes it took to get started at all. The estimate covers the work. Reality includes the wrapper around the work — and for an ADHD brain, the wrapper is often bigger than the work.

Notice that all three are information problems, not effort problems. You can't fix them by caring more. You fix them by getting real numbers.

The Fix Is Calibration, Not Better Guessing

Here's the reframe that makes this tractable: you don't need accurate estimates. You need a known error rate.

A thermometer that always reads five degrees high is still a useful thermometer, as long as you know it reads five degrees high. Your estimates are the same. If you consistently guess 30 minutes for things that take 75, you don't have a broken instrument — you have an uncalibrated one, running at about 2.5x. And once you know the number, you can correct for it in one second of arithmetic, without having to fix your sense of time at all.

The goal of the next two weeks is to find your number.

How to Calibrate in Two Weeks

Step 1: Write the estimate down before you start

Every time you begin a task, commit to a number first, in writing. Not in your head — in your head it will quietly revise itself to match whatever happened, which is exactly how you've stayed uncalibrated this long.

Keep it crude. A note app, a sticky, a column in your task manager. Task name, estimated minutes. That's it.

The discipline that matters: estimate before you start, never after. An estimate written retroactively is worthless as data.

Step 2: Capture the actual — without relying on yourself

This is the step where most people's attempt dies, because manually logging your stop time requires the exact thing ADHD is worst at: noticing that time has passed and interrupting yourself to record it. You'll do it for three days and quit.

So pick a method that doesn't depend on your attention:

• Run a timer for every session. Start it as part of starting, and let the timer own the record. My broader setup for this — timer, sessions, check-ins — is in the Mac setup I use to make time visible.
• Or let it log itself. Automatic activity tracking records which apps and documents you were actually in, so the timeline exists whether or not you remembered to press stop. This is why Focusmo tracks activity locally in the background — the whole point is to get a real record from a brain that makes an unreliable witness.

Either way, you want the actual duration to end up on paper next to the estimate, with as little dependence on your own vigilance as possible.

Step 3: Sort into a handful of task types

Two weeks in, you'll have a messy list. Don't analyze it task by task — that's noise, and it's also where people abandon this because it feels like a spreadsheet project.

Instead, bucket everything into four or five recurring categories that reflect your real work. For example:

• Email and messages
• Writing / drafting
• Meetings and their prep
• Deep technical work (coding, analysis, design)
• Admin (forms, expenses, scheduling)

Four is plenty. The multiplier is almost never the same across categories, and that difference is the most useful thing you'll learn.

Step 4: Do the arithmetic

For each category, total the estimated minutes and total the actual minutes, then divide:

Your multiplier = total actual ÷ total estimated

If you estimated 300 minutes of admin across two weeks and it actually consumed 690, your admin multiplier is 2.3.

Two things usually surprise people here. First, the multipliers are rarely uniform — a common pattern is being roughly accurate on well-defined technical work while being wildly optimistic on admin and communication, because those are the tasks with the biggest invisible wrapper. Second, the numbers are frequently above 2. That's not a moral verdict, it's a calibration constant.

Step 5: Apply it at the moment of commitment

The multiplier is only useful if it fires before you say a number out loud. Build the habit at the point of promising:

Guess honestly → multiply by that category's number → say the second number.

Not the first one. The first one is the one that's been getting you in trouble. If your gut says "an hour" for a writing task and your writing multiplier is 1.8, you block two hours and you tell people two hours.

The relief here is real and a little unexpected: you stop having to defend a running estimate all afternoon, and you start finishing things inside the window you named.

Estimating Something You've Never Done Before

Multipliers work on recurring work. For a genuinely novel task, use decomposition instead:

1. Break it into pieces you have done before. You may not know how long "build the onboarding flow" takes, but you've done "write a spec," "build a form," and "review with the team."
2. Estimate each piece and multiply each one by its own category multiplier.
3. Add an integration tax. The pieces never simply sum — there's handoff, context reloading, and rework between them. Add 20–30% on top and adjust that figure as you learn.

And when the honest answer is "I have no idea," say that instead of inventing a number. "I'll know after an hour of scoping" is a far more professional answer than a confident guess you'll miss by 300%.

Where This Goes Wrong

You track too much and quit. Four categories, two weeks, one number each. If it feels like a research project, you've overbuilt it and you'll abandon it.

You estimate the work, not the whole errand. The estimate must cover start-to-finish including setup, context-switching, and the getting-started friction. If starting is where your time actually disappears, that's a distinct problem with its own fix — see ADHD task paralysis and how to start.

You treat one bad week as your identity. A single derailed project doesn't define your multiplier; totals across two weeks do. Recompute every couple of months, since the number moves as your work changes.

You use the data to beat yourself up. This is the big one. The log will show things you'd rather not see — the hour eaten by Slack, the "quick" task that took a morning. That's the raw material, not an indictment. If reading your own numbers makes you feel like a failure rather than an engineer with a measurement, throttle back to tracking one category.

What Better Estimates Actually Buy You

Once your numbers are honest, a few things get easier fast.

You can plan a day that fits. Most overloaded schedules aren't a discipline failure — they're an arithmetic failure, built by stacking six tasks that were each underestimated by half. If you want to sanity-check how much focused work realistically fits in your day at all, the free Deep Work Calculator is a quick reality check, and the ADHD Focus Planner shapes a schedule around when you actually focus rather than when you wish you did.

You can also say no with evidence. "That's a six-hour task and I have four hours this week" is a conversation you can have with a manager. "I feel like that's a lot" is not.

The Takeaway

You are not bad at time. You are running an uncalibrated instrument and reading it as if it were accurate. Write the estimate down before you start, capture the actual without relying on your own attention, bucket two weeks of data into four categories, divide, and then say the corrected number instead of the raw one.

It takes two weeks of light logging to buy back years of blown deadlines and apologies you didn't deserve to make.

If you'd rather not maintain the log by hand, try Focusmo — it's a free Mac menu-bar app with focus sessions and automatic local activity tracking, so the "actual" half of your calibration records itself while you work.