How to read your own menstrual cycle

The menstrual cycle is counted from the first day of bleeding to the day before the next period starts. The American College of Obstetricians and Gynecologists puts the normal adult length at 21 to 35 days and a normal period at up to 7 days. The Office on Women’s Health uses a narrower window, 24 to 38 days. NICHD gives the average as 28 days and states that ovulation can occur anywhere between 10 and 21 days after day one. In an analysis of 612,613 ovulatory cycles published in npj Digital Medicine, the mean cycle was 29.3 days, with a mean follicular phase of 16.9 days and a mean luteal phase of 12.4 days.

What actually happens in each phase of the cycle?

Four phases, four different hormone profiles, four different sets of measurable effects. Menstruation and the follicular phase run on low then rising estradiol. Ovulation is a luteinising hormone event, and the Office on Women’s Health notes that LH begins surging about 36 hours before the egg is released and peaks about 12 hours before. The luteal phase belongs to progesterone. Each phase produces a physical change you can observe without a lab.

The useful move is to stop thinking of the cycle as one month-long mood and start reading it as four segments with different rules. Temperature, sleep architecture, appetite and pain do not drift randomly. They track the dominant hormone. That is why the same woman can feel physically unrecognisable on day 6 and day 24 without anything being wrong.

Why does body temperature rise after ovulation?

Progesterone acts on the preoptic area of the hypothalamus and raises the temperature at which your body starts trying to cool itself. Resting core temperature sits 0.3 to 0.7 degrees Celsius higher in the luteal phase than the follicular phase. It is not extra heat production alone. It is a moved set point.

Baker and colleagues, reviewing the thermoregulation literature, describe the sequence precisely: core temperature rises about 24 hours after a detectable increase in plasma progesterone and reaches a plateau within 48 hours. The thresholds for sweating onset, cutaneous vasodilation and cold sensation all shift upward at the same time. In animal work, progesterone applied directly to the preoptic area inhibits warm-sensitive neurons, which suppresses heat loss.

The practical consequence is the part period apps rarely state. In a cycle with no ovulation, there is no progesterone rise, so there is no temperature shift at all. The shift is evidence that a corpus luteum formed, not a prediction that one will.

A temperature rise is not a fertility signal. It is a receipt for an event that already happened.

Is a 24 day menstrual cycle normal?

Yes, under every major US definition. ACOG’s threshold for abnormal is shorter than 21 days or longer than 35. The Office on Women’s Health sets its lower bound at 24 days. A 24 day cycle sits inside both. A 19 day cycle sits outside both.

This is the number most often misquoted online, so it is worth naming exactly who says what. ACOG Committee Opinion 651 states that by the third year after menarche, 60 to 80 percent of cycles run 21 to 34 days, as is typical of adults. NICHD says 21 to 35 for most women and 21 to 45 for teenagers. The Office on Women’s Health uses 24 to 38. Checking those four documents against each other, rather than against secondary coverage, is the method we describe in how we research.

Why does the length of my cycle change but not the second half?

Because the two halves are governed by different clocks. The follicular phase waits for a follicle to mature, and that wait stretches or shortens. The luteal phase runs on the lifespan of the corpus luteum, which is far more consistent. Cycle length variation is mostly follicular variation.

The npj Digital Medicine analysis makes this measurable. Across 612,613 cycles, the mean follicular phase was 16.9 days with a 95 percent confidence interval of 10 to 30 days. The mean luteal phase was 12.4 days with an interval of 7 to 17. Between ages 25 and 45, mean follicular phase length fell by 0.19 days per year of age. Luteal phase length showed no significant difference across age cohorts.

So a cycle that used to be 30 days and is now 26 has usually lost follicular days, not luteal ones. That specific pattern is worth understanding on its own, which is why we cover why a cycle starts getting shorter separately.

What does progesterone actually do in the second half?

Progesterone stabilises the endometrium, raises core temperature, and converts in the brain to allopregnanolone. Allopregnanolone is a potent positive allosteric modulator of the GABA-A receptor, with sedative and anxiolytic properties. Progesterone and allopregnanolone both rise through the luteal phase and fall sharply just before bleeding.

That fall is the mechanism behind a large amount of premenstrual symptom research. Hantsoo and Epperson describe chronic exposure followed by rapid withdrawal from ovarian hormones as a key factor in the aetiology of premenstrual dysphoric disorder, with altered GABA-A receptor function in the withdrawal models. Under DSM-5 criteria, PMDD requires at least five symptoms in the week before menses, at least one of them affective, absent in the week after menses.

Appetite tracks the same curve. A systematic review and meta-analysis in Nutrition Reviews pooled 15 datasets covering 330 women and found significantly higher energy intake in the luteal phase, a crude average difference of 168 kilocalories per day. That is a phase effect, not a lapse in willpower.

The sedative reputation of progesterone is where most explanations stop, and it is incomplete. We take that apart in detail in why progesterone makes you sleepy.

Why is sleep worse in the luteal phase if progesterone is sedating?

Because the measurable sleep changes go the other way in midlife. In a polysomnography study of 20 women in the early menopausal transition, mean age 48.8 years, both awakenings and arousals per hour of sleep increased in the luteal phase compared with the follicular phase, and slow wave sleep decreased.

De Zambotti and colleagues confirmed the luteal phase biochemically, using a progesterone level of 3 nanograms per millilitre or higher. The luteal nights also showed greater sleep spindle density, longer spindles and increased 14 to 17 hertz EEG activity. The effect held whether or not the woman had a diagnosed insomnia disorder. Sleep was more disrupted in the luteal phase in both groups.

This is the reconciliation: allopregnanolone is sedating at the receptor, and luteal sleep is still more fragmented on the recording. If you are already tracking broken nights, read that alongside why you keep waking up at night.

Is a short luteal phase something to worry about?

A luteal phase under 10 days meets the clinical definition of luteal phase deficiency used in the NIH BioCycle Study. In that cohort of regularly menstruating women aged 18 to 44, it appeared in 8.9 percent of 463 ovulatory cycles. It recurred across cycles in only 3.4 percent of women.

Schliep and colleagues also tested a biochemical definition, a peak luteal progesterone of 5 nanograms per millilitre or less, which appeared in 8.4 percent of cycles, with only 4.3 percent meeting both criteria. Both definitions were associated with lower follicular and luteal estradiol. Only the short-duration version was associated with lower LH and FSH, which suggests the two are not the same problem.

The number that matters is recurrence, not incidence. One short luteal phase is common. A short one most months is a different observation, and it is the version worth raising with a clinician.

Why do periods hurt, and when is the pain a signal?

Primary dysmenorrhea is caused by prostaglandins made in the uterine lining, which make the muscles and blood vessels of the uterus contract. ACOG states that prostaglandin levels are highest on day one and fall as the lining sheds, which is why pain usually eases after the first days.

ACOG also states that more than half of women who menstruate have some pain for 1 to 2 days each month, and that primary dysmenorrhea often becomes less severe with age. That direction of travel is the useful reference point. Pain that follows the prostaglandin curve, front-loaded and fading, behaves like primary dysmenorrhea.

Pain that arrives before bleeding and outlasts it, pain that worsens year on year, or new pain in a cycle that was previously comfortable does not follow that curve. Secondary dysmenorrhea, caused by a condition in the reproductive organs, is the category that fits a changing pattern.

When does a changing cycle mean perimenopause rather than a bad month?

The STRAW+10 criteria define early menopausal transition by one measurable thing: a persistent difference of 7 days or more in the length of consecutive cycles. Persistence has a definition too. The variable-length cycle must recur within 10 cycles of the first one.

Harlow and colleagues also describe an earlier stage. In STRAW+10 stage minus 3a, cycles remain regular but become subtly shorter, with early follicular phase FSH rising and becoming more variable. The late transition is marked by amenorrhea of 60 days or longer. The NIA states that most women begin the transition between ages 45 and 55, and that the average age of menopause in the United States is 52.

Two cycles seven days apart is not a stage. A seven day swing that keeps coming back within ten cycles is. That distinction is the whole reason we wrote up the perimenopause symptoms nobody warns you about.

How do you read your own cycle as a data series?

You need three columns and a few months, not an app subscription. The point is not to hit a textbook 28 days. It is to establish your own baseline and then detect a change against it, which is the only comparison that carries information.

1. Log day one of every period. Cycle length is day one to the day before the next day one. Everything else is derived from this.
2. Record waking temperature before getting out of bed. You are looking for a sustained shift of about 0.3 degrees Celsius or more, not a single high reading.
3. Count backwards from the next day one to the temperature shift. That is your luteal phase length. Under 10 days, repeatedly, is the finding worth reporting.
4. Note the difference between consecutive cycle lengths. Seven days or more, recurring, is the STRAW+10 early transition threshold.
5. Track symptom timing, not just severity. Whether a symptom is luteal, menstrual or follicular narrows the mechanism faster than how bad it was.

What is the single most useful number to watch?

Not cycle length. Cycle-to-cycle variation. A woman whose cycles run 26, 27 and 26 days is more regular than a woman averaging a textbook 28 across 22, 34 and 28. The average hides exactly the signal that both ACOG and STRAW+10 use as their threshold.

This is the writer’s view, and it is where the standard advice is weakest. Publishing a single normal range invites people to check whether they are inside it, which is a one-off yes or no. The npj Digital Medicine data found that per-user cycle length variation rose sharply above age 40, reaching 3.1 days at age 45, while mean cycle length fell. Variation was the parameter that moved.

None of this is a diagnosis, and no chart replaces an examination. What a chart does is turn a vague sense that something changed into a dated, quantified observation that a clinician can act on.