I counted every phone pickup for fourteen days, 12–25 October 2026, with no intention of changing anything. Published: 28 October 2026 · Last updated: 28 October 2026 · Reading time: ~9 minutes · The tally, the photographs and the notes in this post are my own.
Every experiment on this blog works the same way. I change one thing, I write it down, I report what happened. Seventeen times now. This is the eighteenth, and it is the first one where the counting was the whole experiment. Nothing switched off, nothing added, nothing to change but the pencil.
For fourteen days I counted every time I picked up my phone, and I made no attempt to reduce it. No app blocker, no rule, no target. Just a tally on a piece of paper, one mark per pickup, and the intention to be accurate.
It fell by 31%. Over two weeks. Without me trying.
The rule, and what a pickup is
One mark every time the phone came off a surface. Not every time I unlocked it. Every time I picked it up, including the times I looked at it and put it back down. If it was in my hand, it counted, whether or not the screen woke up.
Three exceptions: answering a call while it was already in my hand did not count, using it as a satnav did not count, and neither did setting an alarm at night. Those were the only three I allowed, and I allowed them on day one, before I knew what the numbers would be.
The tally was a piece of paper on the kitchen counter with fourteen rows. Marks in pencil. In the evening, I counted the marks and wrote the total. That is the whole method.
The fourteen days
| Day | Pickups | Note |
|---|---|---|
| 1 | 62 | Started counting at lunchtime and had to estimate the morning. |
| 2 | 58 | Nothing unusual. Noticed I pick it up to check the time and then do three other things. |
| 3 | 60 | Worst day. Waiting room for forty minutes and no book. |
| 4 | 55 | Left it in the kitchen for a stretch in the afternoon without meaning to. |
| 5 | 57 | Caught a pickup that lasted about two seconds. Decided it still counted. |
| 6 | 53 | Saturday. More pickups than I expected, mostly out of boredom rather than need. |
| 7 | 55 | Sunday. Counted eleven before nine in the morning, which surprised me. |
| Week one average: 57.1 | ||
| 8 | 51 | Back at work. Fewer, because I was in meetings most of the day. |
| 9 | 48 | Went to pick it up, saw the sheet, put it back down. Did not feel like a decision. |
| 10 | 50 | A bump. Bad news in the morning and I checked it every twenty minutes. |
| 11 | 46 | First day under fifty. Did not notice at the time. |
| 12 | 44 | Left the phone in the hallway by accident for two hours. |
| 13 | 47 | Saturday again, but lower than last Saturday. No idea why. |
| 14 | 43 | Lowest day. Saw the number in the evening and did not feel like I had tried. |
| Week two average: 47.0 · fourteen-day average: 52.1 | ||
Day one was 62, day fourteen was 43. That is nineteen fewer pickups, a drop of about 31%, and I want to be clear that I did not decide to do this. There was no day where I looked at the sheet and thought "I should cut down." The sheet was on the counter and the phone was in my pocket, and the number went down anyway.
Why I think it happened
I have three explanations and I cannot separate them, so I am giving you all three and being honest that this is where the evidence runs out.
The first is that a tally is a decision point. On day one, picking up the phone was a single action. On day two it was two: pick it up, make a mark. The mark is a small cost, and some fraction of pickups were not worth the cost. I did not decide not to pick it up. I decided not to make the mark, and the pickup did not happen. Day nine is the clearest example in the log, and the note I wrote that evening does not mention willpower at all. It says I put the phone back down and it did not feel like a decision.
The second is that I was watching. Not trying to change, but paying attention in a way I normally do not. Attention is not neutral. If you spend a fortnight noticing something, you will notice it more, and the noticing is not free of consequences.
The third is that the sheet was on the counter. This is the least interesting explanation and the one I would bet on. A piece of paper on the kitchen counter, in pencil, with marks on it, is a physical object in my flat that did not exist before. It changed the room. I do not think I can claim any credit for the drop without admitting that I changed the furniture. Day twelve, where the number fell because I left the phone in the hallway by accident, is the day I stopped being sure this was about me at all.
The uncomfortable part
If counting alone can reduce a behaviour by 31%, then every experiment on this blog is contaminated.
In the cleaning experiment, I logged thirty days and the routine stuck. Was it the routine, or the log? In the phone charger experiment, I recorded each morning and the phone-first mornings fell. Was it the charger being in the hallway, or the fact that I was writing down whether it worked? In the coffee experiment, I scored the difficulty out of ten every night and the scores fell. Was that the caffeine leaving my system, or the scoring?
I cannot answer that from any of those posts. In every one of them, the intervention and the record arrived on the same day, and I never separated them.
This is the classic version of the problem, and it has a name I am not going to use because it makes it sound solved. It is not solved here. What I have done, seventeen times, is run an experiment on myself in which I was simultaneously the subject and the observer, and reported the result as though the observer had no effect. That is not a method. It is a method plus a confound, and I have been writing up the method.
Which does not mean the seventeen experiments are wrong. The cleaning routine probably did help, and the charger probably did change my mornings, and the coffee difficulty probably did fall because of the caffeine. The point is that I cannot tell you how much of each was the thing and how much was the pencil.
And there is one experiment in the archive that is closer to a control than I recognised at the time. The wardrobe log ran for four weeks with no intervention, and I ticked boxes and changed nothing. But it counted a different thing from this one. It counted presence, not frequency: whether each item was worn at all, not how many times. And that difference matters, because presence gives you no feedback between days. Once you have ticked the shirt, you have ticked the shirt. Wearing it again adds nothing. Counting a rate gives you a number that changes every evening, and a number that changes every evening is a number you can react to. Counting a binary does not. You cannot count your way into wearing fewer shirts, because the count has nothing to tell you until the four weeks are up. And you can count your way out of picking up a phone, because the count is different every night.
The thing I got wrong
I assumed that the record was a neutral instrument. It is not. The record is an intervention.
That reframing changes what these posts are. If the log is doing part of the work, then "I ran this experiment and it worked" is not quite the right claim. The right claim is: "I started paying attention to this, and this is what happened." Those sound similar and they are different, and the second one is true.
It also gives a better answer to the question people ask, which is "would this work for me?" The honest answer is: it will work if you also write it down, and it may not work if you do not, and I have no idea how much of the effect is the writing. That is a worse answer than the one I have been giving since the first post, and I would rather give the worse answer and have it be correct.
So the rule I came out with is: if you want to change a behaviour, start by counting it and nothing else. Not a target, not a rule, not a reward. Just a number. If it moves on its own, you may not need the rest of the plan. And if it does not move, the plan has to do all the work, and you will know that it is the plan.
What I would do differently
I would have counted first, before any of the seventeen experiments. A fortnight of baseline, with no intervention, would have told me what the pencil alone was worth, and then every later experiment could have been measured against it.
And I would have counted something I genuinely did not care about. I picked the phone because it was easy to count, and it turned out I cared about it, which muddies everything. If I had counted, say, how many times I opened the fridge, the number would probably have moved less, and that would have been a cleaner test of whether counting itself changes things.
Should you do this?
Yes, and it is the cheapest experiment on this blog. Pick one behaviour. Put a piece of paper somewhere you will see it. Make a mark every time the behaviour happens. Do not set a target, do not change anything, do not tell yourself you are cutting down.
Two weeks. Then count the marks.
If the number has fallen, you have learned something about yourself that none of my seventeen experiments can tell you: that attention alone does part of the work, and that a plan built on top of attention may be doing less than you think. If the number has not fallen, you have also learned something, and it is more useful. It means the behaviour is not waiting to be noticed. It is waiting to be changed.
What I cannot tell you is whether the number will fall for you. Mine fell by 31%, and I still do not know why.
Frequently asked questions
Did you cheat?
Possibly on day one, where I had to estimate the morning because I started at lunchtime. That estimate is probably high, which makes the drop look bigger than it is. If I exclude day one entirely, the first day I have full data for is 58 and the last is 43, which is a 26% drop rather than 31%. Still a drop, and still unexplained.
Could you not just have changed the number without noticing?
Yes, and that is the other weakness. I was counting by hand, in pencil, and I wanted the experiment to be interesting. A tally counted by the person whose behaviour is being counted is not a reliable instrument. An app would have been better, and I did not use one because an app that tracks pickups also tells you what to do about them, which is exactly the intervention I was trying to avoid.
Would a phone app be better?
For accuracy, yes. For this experiment, no. Every app I looked at has a target, a streak, or a notification, and all three are interventions. If you want a clean count, use a sheet of paper. If you want to reduce your phone use, use the app. They are different jobs.
Would you do it again?
The counting, yes, and I have. I now do a fortnight of baseline before any new experiment, which is the change this made to how I work. What I would not repeat is seventeen experiments without one. That is the thing this post taught me, and it arrived at experiment eighteen, which is later than it should have.
About this post
Written by Maha Alkumaish. I write about the routines and systems that survive a real working week, tested on myself before they get written about. The tally sheet and the photographs in this post are my own.
All photographs in this post are my own. I was not paid or sponsored by any brand mentioned, and there are no affiliate links in this post. More about me and how these experiments are run.
Related reading on this blog:
- I moved my phone charger out of the bedroom, the same behaviour, changed by moving an object instead of by counting, and I can no longer tell you which did the work
- I stopped drinking coffee for two weeks, where I scored the difficulty out of ten every night, and the scoring is now a suspect
- I logged what I actually wore for four weeks, the control I did not know I had, where counting could not move a behaviour that was not optional
- More Wellness & Routines experiments
0 Comments