The strange part is how long I have known about that breath. I have known about it for years. I have known about tapping for years, and about the butterfly hug on your chest to come back when you have gone somewhere far away. I read about all of it. I understood the mechanisms well enough to explain them to someone else over dinner. And in years of hard meetings I had never once used any of them. The information was in my head, and it was sitting there doing nothing.

What changed was not that I read something better. It was that for the last few months I have been turning the dial most days, answering which of the four states I am in, and then doing the one small thing it offered. Easy long exhale: in for four, out for six. Butterfly hug: arms crossed, tapping my own collarbones. Feet on the ground: press down, feel the chair hold you. One minute, sometimes five. And then, which I think is the part that actually mattered, I noticed what it did. Not whether it was supposed to work. Whether my shoulders had come down. Whether my jaw had let go. Whether the fizzing in my chest was still fizzing.

Two smaller things did more than I expected. When something worked I starred it, which moves it to the front of what I am offered next time, so the list slowly stopped being a general list and became mine. And when my therapist gave me something the app did not have, I typed it in myself. Three of the tools I reach for most often are ones I added. The toolkit I am using now is not the one I shipped.

When I went looking afterwards, the research said my years of knowing and not doing were entirely ordinary. In 2006, Thomas Webb and Paschal Sheeran pooled 47 experiments in which people were randomly assigned to something designed to make them intend to change a behaviour. A medium-to-large shift in what people intended, d = 0.66, produced only a small-to-medium shift in what they actually did, d = 0.36. (Webb and Sheeran, 2006, Psychological Bulletin) Being convinced moves you about half as far as the conviction feels like it should. And I had not even been convinced by an intervention designed by researchers. I had been convinced by a book.

What does close the gap is less interesting and much slower. In 2010, Phillippa Lally and colleagues at University College London asked 96 volunteers to pick one small eating, drinking, or activity behaviour and do it every day, in the same context, for 12 weeks. The volunteers rated how automatic it felt each day. Automaticity rose along a curve, and the time it took people to get within reach of their own ceiling ranged from 18 days to 254. The authors fitted their model to 62 of the volunteers and got a good fit for 39, so this is a small study making a modest claim, and the claim is the useful part: automaticity comes from repeating a behaviour in a consistent context. Missing a single day did not meaningfully set anyone back, but being inconsistent did. (Lally et al., 2010, European Journal of Social Psychology)

There is one more piece, and it is the one that explains the noticing. In 2017, Kaitlin Woolley and Ayelet Fishbach ran five studies on why people stay with long-term goals. People say they exercise or study for the delayed payoff, and the delayed payoff is what got them to set the goal. It was not what predicted whether they kept going. Immediate rewards, such as simply enjoying the thing while doing it, predicted persistence where delayed rewards did not, over a week and over three months. (Woolley and Fishbach, 2017, Personality and Social Psychology Bulletin) Reading about a practice hands you the delayed reason to do it. Feeling your shoulders drop hands you the immediate one. For years I had the reason and never the reward.

James Clear tied this exceptionally well in his book Atomic Habits. His two-minute rule says that “when you start a new habit, it should take less than two minutes to do,” which is almost exactly the shape of what I ended up doing, and the fourth of his four laws is to make a habit satisfying, on the argument that what gets rewarded immediately is what gets repeated. (Clear, How to Stop Procrastinating) He had also written up the Lally study years before I went looking for it, although he gives the median as “66 days to be exact”, and the paper is a good deal less certain than that. (Clear, How Long Does it Actually Take to Form a New Habit?) So the one-minute practice and the immediate reward had both been described to me, in plain English, in a bestseller, years ago. It changed nothing, because reading the book about why reading is not enough is still reading.

Here is what I believe now. Knowing a practice and having one are two different things, and the distance between them is not more information. It is a small number of minutes, done often, in a place your body comes to expect them, with your attention on what changed in your body afterwards rather than on whether it worked. I did not get these techniques from understanding them. I got them from doing them one minute at a time, for a few months, while paying attention.

I built this to find out whether it works for people who are not me, and I still do not know that. What I can report is that it worked on me, surprisingly well, and that it did not arrive as an insight. It arrived as something my body did in a meeting before I got there.

If you have a practice that eventually became automatic for you, I would like to know what the turning point was, because I no longer think mine was the moment I understood it.

I wear two trackers. Yes, two, and I sleep in both of them. Last night one of them reported that I had been awake for two hours in the middle of the night. The other failed to pair and had nothing to say. On most mornings the two devices and my own sense of things form a small committee, and the devices get the deciding vote. This morning the committee was short a member.

So I did the thing I have spent a year building an app to help people do. I stopped and asked my body a direct question: am I actually tired, or do I feel like someone who was told she slept five hours? I sat with it for about thirty seconds. The answer was that I was rested. I was not wired and I was not foggy. I was rested.

I went to morning yoga. I am fairly sure that if the second tracker had paired and agreed with the first, I would have skipped it, because two devices saying you were up for two hours is very persuasive at six in the morning. Instead I am more awake and more settled than I have been most mornings this month, and I owe that to a question that took half a minute.

Afterward I went looking for what the research says about the number I almost believed, and it changed how I read these reports.

The part of a sleep report I had been treating as fact is the part the devices are worst at, and this is true of the current generation, not just the early ones. In 2025, a team at Hasselt University and Antwerp University Hospital tested six wrist trackers on sale today, among them the Apple Watch Series 8, the Whoop 4.0, two Fitbits, a Garmin, and a Withings, against polysomnography, the laboratory standard. Every device recognized sleep more than 90 percent of the time. At correctly recognizing wake, the best of them managed 52 percent and the worst 29 percent, and every one of them underestimated the time a person spent awake during the night, by 12 to 48 minutes. (Schyvens et al., 2025, SLEEP Advances) A 2024 study at Brigham and Women’s Hospital, funded by Oura, found the same shape with the Oura Ring Gen3, the Fitbit Sense 2, and the Apple Watch Series 8: at least 95 percent on sleep, and between 52 and 69 percent on wake. (Robbins et al., 2024, Sensors) The field does not even agree on which direction the error runs. When a 2025 meta-analysis pooled 24 studies and 798 people, wrist trackers as a group overestimated night-time wake by about 13 minutes. (Lee et al., 2025, Journal of Clinical Sleep Medicine) Some devices miss wake, some invent it, and the number I was about to plan my morning around is the least settled number on the screen.

Sleep clinicians have a name for what I was doing. In 2017, Kelly Baron and colleagues at Northwestern published a paper in the Journal of Clinical Sleep Medicine about patients who arrived seeking treatment for sleep problems they had diagnosed from their tracker data. They called it orthosomnia. One patient, after a night in the sleep lab showed she was getting deep sleep, asked, “Then why does my Fitbit say I am sleeping poorly?” The authors wrote that consumer trackers “are unable to accurately discriminate stages of sleep and have poor accuracy in detecting wake after sleep onset.” (Baron et al., 2017, Journal of Clinical Sleep Medicine)

And being told you slept badly does something on its own. In 2014, Christina Draganich and Kristi Erdal ran an experiment with 164 students. Each student was told, by what looked like a measurement, either that they had spent 28.7 percent of the night in REM sleep, which is above average, or 16.2 percent, which is well below. The numbers were invented. The students who were told they had slept poorly did worse on a test of attention and information processing, and their own sense of how they had slept did not predict their score. The researchers called it placebo sleep. (Draganich and Erdal, 2014, Journal of Experimental Psychology: Learning, Memory, and Cognition)

Put those together and the morning makes sense. A device with a poor record on wake told me I had been awake for two hours, and if I had believed it, the belief alone could have made me perform like someone who had. The only instrument in the room with a good record on how I actually felt was me, and it had not been asked.

Here is what I believe now. The numbers are a second opinion, and my body gets the first one. I still wear the trackers, and I still look at them. Their measurements are useful for what they are good at, which is patterns over weeks. They are not good at telling me how I am this morning, and I had been letting them answer that question for years.

If you wear a tracker, I would like to know what happens on the mornings when it disagrees with you. Do you check the number before you check yourself, and which one wins?

I had every number. I checked my HRV each morning. I knew my sleep stages, my deep-sleep minutes, and my cumulative stress score. I was one of the most measured people I knew, and I had no idea how I actually was. The data told me what my body had done. It never told me what to do next.

And I was not naive about any of this. I had been in therapy for years. I had read enough about the vagus nerve to award myself a small, self-declared PhD in it. I had studied the brain the way other people study markets: why it breaks, what motivates it, why we sleep. I knew the theory cold. And with five deadlines and two young kids, I could not see that I was in a meltdown. Knowing how the nervous system works turned out to be no protection from my own.

Therapy was different. Once a week, my therapist gave me something practical I could do in one to five minutes. Those tools worked. The problem was where they lived: on post-its, in the Notes app on my laptop, in a journal at home. When I needed them most, I could not find them.

Before I went back to work, my anxiety came down to one question: would I hit the bottom again? So my therapist, Aubrey Landau, LMFT, gave me a framework. Four states: green, yellow, red, and purple. The first step is not to fix anything. The first step is to answer honestly which one you are in right now.

I tried to do this on my phone, and I learned something. Picking up my phone to check in with myself meant I checked my email, answered two texts, opened Instagram, and looked at my HRV. I did everything except check in.

So I built something else. Solastone is a stone you turn to say where you are, and it then offers one small practice that fits the minutes you have. The app is out first, because I need to know whether this works for people who are not me. The physical stone is next.

Here is what I believe after this year. Measuring your body is not the same as being in it. The tools that help are small, and they only help if they are within reach. And asking “where am I right now?” is a skill, not a feeling. It gets better with practice, and it is worth practicing before the bottom, not after.

If you have been through a version of this, I would like to hear what worked for you. That is how the toolkit in Solastone grew, and it is how it will keep growing.