Habit tracking with ADHD: fewer habits, louder feedback

Most habit advice assumes the hard part is deciding. With ADHD the deciding is easy and often enthusiastic: twelve habits, colour coded, starting Monday. The hard part is Wednesday.

This is a guide about designing a tracking setup that survives Wednesday. It comes from my own logs and from a lot of email with people who tried the same thing. It is not medical advice, and nothing in it replaces treatment or a therapist. It is about the mechanics of the tool.

The failure is not motivation, it is the gap

Two gaps break most setups.

The first is the gap between the intention and the cue. You decide to stretch every evening. There is no moment in your evening that reliably says “now”. Without that moment the habit depends on remembering, and remembering is the thing that is unreliable.

The second is the gap between doing and seeing. You did the thing. The app registers it silently. Nothing happens. A brain that runs on immediate feedback gets nothing back, so the loop never closes, and by week three the app is a list of accusations you scroll past.

Every recommendation below is aimed at one of those two gaps.

Cut to four habits, and cut the right ones

Four is not a magic number, it is a ceiling that survives a bad day. Twelve habits on a good week becomes twelve reasons to feel behind on a bad one, and the bad ones are the weeks that decide whether the app stays installed.

When you cut, keep habits that have a physical trigger already attached. Teeth after breakfast. Medication next to the kettle. Water bottle refill when you sit down at the desk. Those already have a “now” in them. Habits with no natural anchor (“read more”, “be present”) need one built, and building anchors is slow work. Do it for one habit at a time.

Cut anything you are tracking to feel virtuous rather than because the number matters. A tracker full of aspirational entries is a mood board, and mood boards do not survive contact with a Wednesday.

What I ended up with, after a year of trimming: medication, water, deep work minutes, and steps from Health. Two need one tap. One needs a timer button. One needs nothing at all.

Make the app do the work you cannot

The single biggest change is moving habits from “you remember and tap” to “your phone already knows”.

Steps, sleep, workouts and stand hours are recorded whether you engage or not. Reading them from HealthKit means the habit is logged with no interaction. That is not cheating. The purpose of the row is to know, and knowing without an action is strictly better.

The second lever is the widget. An interactive Home Screen widget lets you check a habit from the place your thumb already goes twenty times an hour, with no app launch and no sync round trip. The distance between intent and completion collapses to one tap in the place you were already looking.

The third is the timer with a Live Activity. When a 90 minute deep work block runs, the countdown sits on the Lock Screen and in the Dynamic Island. It is not a reminder to start, which you would ignore. It is proof that you already started, which is much harder to walk away from. Time blindness is easier to argue with when the time is visible.

Louder feedback, not more notifications

Notifications are the obvious answer and they are usually the wrong one. A notification you ignore teaches you to ignore notifications, and after a fortnight the app is background noise with a red badge.

Feedback is different. Feedback is what happens in the two seconds after you log something.

What works, in rough order:

A value that moves. 6/8 becoming 7/8 is more satisfying than a checkbox turning green, because a number implies a distance and closing a distance is legible.

A grid that fills. The year heatmap is the slowest and most durable reward in the app. A square a day is not exciting on Tuesday. Ninety squares is unmistakable, and unlike a streak counter it survives a hole.

A shield banking. Overshoot your target for five days in a row and you bank a shield. That produces a small, concrete reward for consistency that is not simply “the number went up by one”.

What does not work: confetti, badges, a character who is sad, and any message that implies disappointment. If an app can be disappointed in you, it will eventually be deleted for it.

Streak shields, and why they matter more than streaks

Here is the mechanic, plainly. Every five consecutive days you bank one shield, up to a maximum of three. Miss a day and a shield spends itself: the chain holds, and the log still shows the day as missed. Nothing is rewritten and nothing is faked.

Why this matters more with ADHD than without it: the classic streak is a system where one bad day destroys three months of evidence. If your relationship with consistency is already fraught, that mechanic is not a motivator, it is a trap with a countdown on it. People do not slowly drift away from a broken streak. They delete the app the same evening.

The shield changes what a missed day means. It is a cost, not a catastrophe. You have three, they regenerate, and the honest record underneath is unchanged. When the shields run out, the chain does break, and that is fine: the point is not to make missing free, it is to stop one flight or one flu from erasing the whole record.

Break mode matters as much. It freezes every habit and spends no shields, so a week away is a frozen week rather than seven failures. The days come back marked frozen, which is information rather than a verdict, and that distinction is most of the difference between a tracker you keep and one you delete in March.

One block instead of five decisions

A morning block groups four habits behind a single trigger and a single tap. This is the structural change that made the biggest difference to me, and the reason is not efficiency. It is that a block is one decision.

Five separate rows means five moments where you can stall. A block means the sequence is already decided: water, medication, stretch, journal, and you tap once when the sequence is done. The decision was made when you set the block up, on a day when deciding was cheap.

Anchor the block to something that already happens. Not 06:30, because 06:30 is a hope. After the kettle boils. After the shower. The trigger has to be an event you cannot skip, otherwise you have just moved the problem.

Notifications worth keeping, and the rest

Keep two kinds. Delete everything else.

Keep the one attached to your anchor, if the anchor is time-shaped rather than event-shaped. A single notification at the moment the block should start is a cue. Six notifications spread through the day are wallpaper.

Keep a weekly summary. Once a week, a local notification that says what the week looked like is worth more than seven daily nudges, because it arrives when you can actually act on it. In miniHabits that is the Sunday digest, and it is generated on the device rather than sent from a server.

Delete streak warnings. “Your 40 day streak is at risk” is a threat, and threats work exactly once. Delete re-engagement pushes. If an app is messaging you to say it misses you, it has stopped being a tool.

Two setups that did not work for me

Worth writing down, because both sounded good on paper.

The colour-coded twelve. Twelve habits, four categories, colours, a beautiful screenshot. It lasted eleven days. Failure mode: every morning began with reading a list and choosing, and choosing is the expensive operation. By day nine I was checking the easy ones and skipping past the rest, which meant the data was actively misleading.

Habits with no anchor at all. “Stretch at some point”, “read more”, “be present”. These are intentions rather than habits, and a tracker turns them into a daily reminder that you are not the person you intended to be. If a row cannot answer the question “after what?”, it is not ready to be tracked. Park it, and pick it up when you know where in the day it goes.

There is a third one I keep half-doing: tracking something because the number is interesting rather than because I intend to change it. That is fine as long as the habit is automatic, like steps from Health, and a problem as soon as it needs a tap.

Medication, specifically

If you track one thing, track this, and give it the least friction available.

Put it in a widget so it takes one tap from the Home Screen. Anchor it to something physical that already happens at the right time, usually the kettle or the toothbrush. Do not put it in a block with three other habits, because the block gets checked at the end of a sequence and this needs to be logged at the moment it happens.

The reason to log it is not discipline, it is the question “did I already take it?” at 09:40 on a morning you cannot reconstruct. A dated entry answers that in half a second, and the pill organiser answers it too, which is why plenty of people are better served by a box from the pharmacy than by an app. Use whichever one you will actually touch.

The first week

Set up two habits, not four. The block comes later.

Pick the two that already have anchors and let the app be almost empty for seven days. An empty looking tracker feels wrong and is exactly right: you are testing whether the cue fires, not building a system.

At the end of the week, look at the time-of-day chart. If both habits happened at roughly the same time each day, the anchor is real and you can add the third and fourth. If the times are scattered, change the anchor rather than adding anything. Adding habits to a setup that is not firing is how the twelve-habit list gets built.

Backfill without guilt, and read the pattern

You will forget to log things. This is not a moral failure, it is the same attention problem wearing a different hat.

Retroactive entry matters for exactly this reason: you can log last Tuesday and the completion maths stays correct. An app that refuses backfill is telling you your data must be wrong rather than late.

Then, once a month, read the patterns rather than the totals. Two views repay the time.

Weekday bias shows completion by day of week. Mine collapses at weekends, by 18%, because every anchor I use is attached to a workday. That is not a discipline problem, it is a missing cue, and the fix is a different weekend anchor rather than more effort.

Time of day shows when things actually happen. Mine peaks at 10:00 and again at 22:00, which told me that scheduling anything at 15:00 was optimism.

Correlations show which habits carry others. Reading is 34% more likely on days I meditated first. That makes meditation load bearing, and it moves to the top of the block.

None of these are visible in a streak counter. They are the reason to keep a record at all.

A setup that survives a bad month

Four habits. Two of them automatic through Health or a widget. One block in the morning with a physical trigger. Shields on. Break mode used without hesitation. One monthly look at the weekday and correlation charts, and nothing else.

When a bad month arrives, and it will, cut to one habit rather than pausing everything. One row, kept alive, is the thing you rebuild from. An empty app is a decision to start over, and starting over is the expensive part.

What miniHabits does about all this

Everything above is a description of the app, so it is fair to say it plainly. Five tracking modes mean the habit picks the interaction: checkbox, counter, timer with a Live Activity, quantity with a unit, and HealthKit for the ones your phone already logs. Shields and break mode are the mercy layer. Blocks group habits behind one trigger and one tap. Interactive widgets remove the app launch. The statistics view carries weekday bias, time of day and correlations.

There is no account, no feed and nothing that can be disappointed in you. The free tier covers ten habits and every tracking mode, which is more than the four this guide recommends.

Two settings are worth changing on day one. Turn the weekly digest on, because a Sunday summary is the review this guide asks for and you will not remember to do it manually. And put the widget on the first Home Screen page rather than the second, because a widget you have to swipe to is a widget you will not use.

One thing the app deliberately does not do: pester you. There is no push campaign, no streak warning, and no notification that fires because you have not opened it in three days. If the app goes quiet it is because you did, and the record will still be there when you come back.

If the gamified end of the shelf is what has worked for you before, the Habitica comparison is honest about what a party does that a private tracker cannot. If what you need is warmth rather than data, read the Finch comparison first. Neither of those is a consolation prize. The right tracker is the one still installed in March.