A period tracker for irregular periods

What counts as irregular, why most apps struggle with it, and how Bloom predicts when your cycle won’t sit still.

Last reviewed 7 October 2026Checked against the sources listed below

If your cycle moves around, most period apps let you down in the same two ways. They predict from an average your body doesn’t keep, and when you’re late they quietly slide the date forward as if nothing happened. This page explains what “irregular” actually means, why prediction is hard when cycles vary, and exactly how Bloom handles it.

What counts as an irregular period

The NHS describes periods as irregular when the gap between them is shorter than 21 days or longer than 35 days.^[1] The U.S. National Institutes of Health gives the same typical range of 21 to 35 days, and estimates that 14% to 25% of women have irregular cycles — shorter or longer than usual, heavier or lighter, or with other problems such as painful cramps.^[2]

Even “regular” cycles are rarely textbook. In an analysis of more than 600,000 cycles logged in a fertility app, the average cycle was 29.3 days and only about 13% of cycles were exactly 28 days long.^[3] The same study found that most of the variation comes from the first half of the cycle, before ovulation, which can run anywhere from about 10 to 30 days.^[3]

Common reasons for a cycle to change include puberty, the years before menopause, pregnancy, starting or stopping hormonal contraception, big changes in weight, stress, and very intense exercise. Conditions such as polyendocrine metabolic ovarian syndrome (PMOS, previously called PCOS) and an underactive thyroid can also cause irregular periods.^[1]

Why period apps struggle with irregular cycles

A prediction is only a guess about the future based on your past. When your cycles are similar, the guess is good. When one month is 27 days and the next is 41, any single number will be wrong a lot of the time. Two design choices make it worse:

• Averaging everything you ever logged. One unusual cycle drags the average, and a body that has changed (after a birth, or coming off the pill) keeps being predicted from cycles it no longer has.
• Rolling into a new cycle when you’re late. If an app assumes a new cycle started on schedule, its phases, its “fertile days” and its next prediction are all built on a period that never happened.

How Bloom handles a cycle that varies

Bloom was designed around these problems. This is what it actually does with the dates you log:

• It predicts from your recent cycles, not your whole history. Bloom uses the median of your last six cycles. A median is not pulled around by one very long or very short month the way an average is.
• Long cycles count. Bloom keeps any gap between periods from 15 up to 90 days as a real cycle, so long cycles are part of your pattern instead of being thrown away as errors.
• Your first guess fades out. When you join, you give a rough cycle length to start from. Each period you log moves the prediction toward your real pattern, and by your sixth logged cycle the starting guess no longer counts at all.
• “Late” is measured against your own normal. Bloom learns how much your cycles usually vary and only calls a period late once it is outside that spread — a couple of days for someone very regular, up to two weeks for someone whose cycles naturally swing. Until it has three cycles to learn from, it allows three days.
• It waits instead of guessing. When your period is due and hasn’t come, Bloom stays where you are and says it is waiting, rather than starting a cycle that hasn’t begun. If you are more than 45 days past due, it stops counting and asks you to update your log.
• It says when it is estimating. Until you have logged three cycles, Bloom labels its prediction as an estimate and tells you how many more cycles it needs. After that it shows the rhythm it has observed in your own logs.

What Bloom can’t do

No app can make an irregular cycle predictable, and Bloom doesn’t pretend to. Its dates stay estimates. It can’t tell you why your cycle changed, and it can’t confirm whether or when you ovulated. It estimates ovulation from your cycle dates, and if you log ovulation tests or cervical fluid it uses those signs to refine the estimate, but a sign you log is a clue, not a confirmation. Irregular cycles sometimes happen without ovulation at all.^[2] Bloom is not a method of contraception, and its estimates should never be used to avoid or plan a pregnancy.

How to track an irregular cycle well

• Log the first day of bleeding every time. It is the single most useful date for you and for a doctor.
• Log when it ends. Period length matters too, especially if periods start lasting longer than a week.
• Add what else was going on. Bloom’s check-in lets you note poor sleep, travel, illness, a heavy week, a new medication or a big life change, alongside symptoms and mood. Months later, that context is often the explanation.
• Bring your history to appointments. Your cycle lengths and symptoms are on your phone in Bloom, so you can show them when a doctor asks how long this has been going on.

If your cycles are irregular because of PMOS (PCOS), there is more on what a tracker can and can’t do in tracking your cycle with PCOS.

Bloom is a wellness and planning app. It is not a medical device, it does not diagnose or treat any condition, and it is not a method of contraception. This page is general information, not medical advice. If something about your cycle worries you, talk to a doctor, nurse or midwife.