What is worth tracking, and what is just noise

There is a version of taking your health seriously that involves measuring almost everything, and there is a version that involves measuring four things. The second one works better, and it is not because the first is too much effort.

It is because most of what you can measure about your own body moves so much on its own that the movement drowns out anything you did.

The one idea that sorts everything

Every measurement has two parts. There is the thing you actually want to know about, and there is the wobble around it: the ordinary variation that happens regardless of your behaviour.

A measurement is worth taking only when the change you care about is bigger than the wobble. If your weight swings by a kilogram and a half in a normal week for reasons that have nothing to do with your eating, then a real half kilogram change is invisible inside that swing. You will see the swings, interpret them as results, and change your behaviour based on nothing at all.

That is the whole test. Not is this measurable, but is the signal bigger than the noise.

What genuinely carries information

A short list passes the test. It is shorter than you would expect and none of it requires a device.

Your cycle length and pattern. This is the single most informative thing a menstruating woman can record, and it takes about four seconds a month. Start date, and roughly how heavy the first two days were. Over six months this shows you the shape of your cycles, which is the earliest and clearest signal that something is shifting, whether that is perimenopause beginning or a thyroid problem developing or nothing at all.^{1 3}

Symptom timing rather than symptom intensity. How bad something felt is unreliable in retrospect and depends heavily on the day you are having. When it happened, relative to your cycle, is reliable and it is the part that carries meaning. A headache on cycle day 24 every month tells you something.² A headache rated seven out of ten tells you almost nothing next month.

Your actual sleep window. Not your sleep score. The clock time you turned the light off and the clock time you got up. Most people are confidently wrong about this by an hour or more, in the direction that flatters them.^{5 6}

Whichever intake you suspect is low, for four days only. Protein or fibre, usually. Four honest days is enough to know whether there is a gap. After that the tracking has done its job and adds nothing.

Why cycle data outperforms everything else

Your cycle is the closest thing you have to a monthly report on a system that is otherwise invisible.

It is useful for a slightly unglamorous reason: it produces one data point a month, so the noise has nowhere to hide. A daily measurement gives you thirty chances a month to see a pattern that is not there. A monthly one gives you one, and by the time you have six of them, a genuine change is unmistakable.

What is worth noticing is the pattern rather than any single cycle. Cycles that were reliably 28 days and are now ranging between 24 and 40. Bleeding that has become clearly heavier over half a year.⁴ A premenstrual week that has extended from two days into eight. Those are real signals, and they are the ones a doctor can actually use.⁸

Why daily weight misleads almost everyone

Daily weight is the clearest example of noise beating signal, and it is the one most people track.

Your weight on any given morning reflects how much salt you ate yesterday, how much food is still physically inside you, how hydrated you are, whether you slept badly, and where you are in your cycle. Fluid retention across a cycle alone commonly accounts for one to two kilograms of movement.

Meanwhile, an actual change in body composition happens at a rate far slower than that. So on any single morning, the number in front of you is dominated by things you do not care about and barely reflects the thing you do.

The practical consequence is worse than useless information. It is that you make decisions from it. You eat less because the number went up, when the number went up because you had soup. You conclude the change is not working, three weeks before it would have shown.

If you want to use weight at all, take it weekly at the same time of day and only ever read the direction across a month.

What wearables measure well and what they estimate

This deserves a fair hearing, because wearables are not fraudulent. They are precise about a few things and speculative about the rest, and the display gives you no way to tell which is which.

What they measure reasonably well is movement, heart rate, and the timing of when you were still and when you were not. That last one gives you your sleep window, which is genuinely useful and hard to get otherwise.

What they estimate is nearly everything that gets top billing. Sleep stages are inferred from movement and heart rate rather than measured, and consumer devices disagree with each other and with clinical measurement on the details.⁷ A single daily readiness or recovery score compresses several estimates into one number, then presents it with a confidence the underlying data does not support.

The metrics that look scientific and mostly are not

Daily step counts as a health verdict. Steps are fine as a rough nudge. The specific number, and the daily target attached to it, come from a marketing campaign rather than from research. Movement across a week is what matters.¹⁵

Basal body temperature for anything other than conception. It is influenced by illness, alcohol, sleep, room temperature and the time you woke. As a general hormonal readout it is far too noisy to interpret.

Continuous glucose monitoring without a diagnosis. Glucose rises after you eat. That is what it is supposed to do. Watching that happen in a healthy woman commonly produces avoidance of ordinary foods based on entirely normal responses.⁹

Frequent at home hormone tests. Hormones swing enormously across a normal cycle and, in perimenopause, from week to week.³ A single reading tells you very little, and a sequence of them mostly measures the swing.

The things worth measuring once rather than repeatedly

There is a separate category that gets confused with tracking, and it is arguably more valuable than any of it.

Some measurements are worth having accurately, occasionally, from a proper test rather than a device. Blood pressure is the clearest example. It is one of the strongest predictors of how the next twenty years go for a woman, it produces no symptoms while it is quietly high, and it is checked in five minutes. Almost nobody logs it, and almost everybody logs their steps.

Ferritin, which reflects your iron stores, is worth knowing if your periods are heavy or you are persistently tired, because low iron is common in women and invisible without a test.¹² Thyroid function is worth checking if your energy, weight or temperature tolerance changed noticeably rather than gradually.¹¹ After menopause, bone density becomes worth discussing, particularly if you have risk factors, because bone loss is silent until something breaks.¹⁰

The pattern here is the opposite of daily tracking. These are slow moving, they are not guessable from how you feel, and they change what someone does for you. That is a better definition of a useful measurement than anything a daily score offers.

Track in bursts, not forever

The most useful pattern is short and deliberate. Pick one question. Track only what answers it. Stop when it is answered.

Four days to find out whether your protein is where you thought. Two weeks to see whether your sleep window is really what you believe. Three cycles to establish whether a symptom is cycle linked or random.

Permanent tracking has a hidden cost that nobody mentions: it turns your attention towards the record and away from how you feel. After a few months, many people can tell you their averages and cannot tell you whether they feel better than they did in spring.

The one thing worth keeping long term

There is an exception, and it is the least technical option available.

A one line note every few days about how you actually felt, dated, is more useful over a year than any dataset. Not a score out of ten. A sentence. Slept badly, snapped at everyone. Good energy all afternoon for once. Bloated again, day 22.

It is more useful for two reasons. It records the thing you actually care about rather than a proxy for it, and it survives being read six months later, which no readiness score does. It is also exactly what makes a doctor’s appointment productive, which is the subject of the next article.

When tracking becomes the problem

At some point measuring can stop being a tool and start being a source of pressure. It is worth naming the signs plainly.

You check the number before you have decided how you feel, and then it decides for you. You feel relief or dread on the way to the scales. You eat, sleep or exercise differently in order to protect a streak or a score rather than because it helps. You avoid an evening out because it will ruin the data.¹³

If any of those are familiar, the correct move is to stop tracking for a month, and to notice that this is not the same as giving up on your health. It is removing an instrument that has started measuring the wrong thing.

When the data is not the answer at all

Tracking is for finding patterns in ordinary variation. It is not a diagnostic tool, and there is a group of situations where more data is the wrong response entirely.

If your bleeding has become genuinely heavy, if you are bleeding between periods or after sex, or if you have any bleeding after menopause, no amount of logging changes what needs to happen next.^{4 3} Those need examining, not measuring.

The same applies to persistent daily bloating, a lump or skin change in a breast, unexplained weight loss, blood in your stool, or low mood lasting more than two weeks.¹⁴ Recording these carefully while waiting for them to resolve is a way of delaying, and it is a common one, because logging feels like doing something.

Track them once, briefly, so you can describe them accurately. Then make the appointment.

The honest limits of this

None of this means your body is unknowable. It means that the things worth knowing about it move slowly, and that the tools which move fast are mostly measuring weather rather than climate.

Four things, tracked in short bursts, will tell you more about your health over a year than a continuous stream of scores. It is also considerably less work, which is convenient, because the version you keep doing is the only version that produces anything.