The Two-Week Sleep Log: What to Record, What to Ignore

Before you change anything about a child’s sleep, spend two weeks writing down six things: when they woke for the day, when each nap started and ended, when the bedtime routine began, when they were actually asleep, when they woke in the night, and what you did about it. That’s the whole log. Everything else — the mood, the dinner, the teeth, the theory about the neighbour’s dog — is noise you will later mistake for a pattern. This is straight Montessori observation, applied to the one part of family life where everyone guesses and nobody records.

Why observation, of all things

Observation is the practice underneath the entire method. A guide doesn’t decide what a child needs and then deliver it; they watch, write down what actually happened, and let the record correct their assumptions. It’s unglamorous and it’s the part that most reliably works.

Sleep is where the gap between assumption and record is widest, because you are collecting data at 2am while sleep-deprived, which is precisely when human pattern-matching is worst. Two weeks of dull, boring entries will tell you things three weeks of theorising cannot — most often that bedtime is later than you believed, or that the “random” waking happens at almost the same time each night.

Two weeks, not two nights. A single bad night is weather. Two weeks is climate.

The six things to record

Keep it to one line per day, on paper, near the bed. If logging takes more than thirty seconds you will stop by day four.

1. Morning wake time. The real one — when they woke, not when you got up.
2. Naps: start and end, each one. Including the eight-minute car nap, which counts more than people think.
3. Routine start time. When bath or pyjamas actually began, not when you meant to begin.
4. Asleep time. When they were genuinely asleep, as best you can tell.
5. Night wakings: the clock time of each. Just the time. Not the duration, not the reason.
6. Your response, in three words. “Walked back twice.” “Fed, resettled.” “Lay down beside.”

That’s it. Six columns, fourteen rows.

The six things to leave out

This is the harder half, and it’s where most sleep logs collapse into a diary nobody can read.

• Reasons and theories. “Probably teeth” is a guess wearing a fact’s clothes, and once it’s written down you’ll believe it.
• Duration of each waking. You cannot measure this accurately half-asleep, and the clock time is the useful variable anyway.
• Feeds, unless you’re specifically investigating night feeds. Otherwise it’s a second data set fighting the first.
• Mood ratings, yours or theirs. They correlate with your exhaustion, not with the child’s night.
• Everything the day contained. Nursery, visitors, the swimming lesson. A fortnight is not long enough to separate those from noise, and the column will tempt you to try.
• What “should” have happened. A log that records the plan alongside the reality becomes a scorecard, and scorecards get abandoned.

The Montessori discipline here is the same one that makes classroom observation work: record behaviour, not interpretation. The interpretation comes later, in daylight, with the whole fortnight in front of you.

Reading the fortnight

At the end of two weeks, read the columns rather than the rows. Four things usually jump out.

Total sleep in 24 hours. Add day and night. This is the number people are most wrong about in both directions.

The gap between the last nap ending and asleep time. This is the single most diagnostic figure in the whole log, and it’s usually the one that’s off.

Whether night wakings cluster. Wakings at roughly the same clock time each night point somewhere quite different from wakings scattered at random.

Whether your response varies. Six different approaches in fourteen nights is the commonest finding, and it’s the one within your control. Any consistent approach beats an inconsistent good one — the argument we made at length in gentle sleep training myths.

What the log usually reveals

Three findings come up again and again, in that order.

Bedtime is later than remembered. The routine starts twenty minutes after intended and takes fifteen minutes longer than assumed, so a 7:00 bedtime is a 7:35 asleep time, five nights out of seven.

The last wake window is wrong for the age. Too short and they’re not tired; too long and they’re overtired, which looks exactly like defiance. If you want an outside reference for what the spacing should be, Betteroo’s guide to nap training is a clear map of how naps and the gap before bed fit together at each age.

The thing you were sure about isn’t in the data. The dog, the dinner, the late bath — most prime suspects simply don’t show up. This is the log’s real gift: it stops you solving problems you don’t have.

When two weeks isn’t the answer

If the fortnight lands in the middle of a developmental leap, the log will look like chaos and the honest conclusion is “not now” rather than a plan. That’s a legitimate result. Betteroo’s explanation of how long a regression actually runs is a reasonable outside read if you’re trying to work out whether to wait or to act.

And a log is not a diagnostic tool. Snoring, mouth-breathing, pauses in breathing, pain on being laid flat, or a child whose total sleep is far outside the normal range for their age are pediatrician questions, not observation questions. Write them down and make the call.

Two weeks of data, now what? Betteroo turns your child's age and your preferred approach into a personalized, gentle day-by-day plan — wake windows, naps and a bedtime that fits what you just observed. Take the 2-minute sleep quiz →

Then change one thing

The observation habit only pays off if the intervention is as disciplined as the record. Change one variable, hold it for a week, and log the same six columns. Move bedtime fifteen minutes earlier — nothing else. Cap the afternoon nap — nothing else. Change three things at once and you have learned nothing except that something worked.

For floor-bed families this matters more than usual, because the room is a variable too: a child who can leave the mattress adds a behaviour to every column. The two-week log is how we’d suggest approaching it in floor beds and independent sleep, and the same observe-then-change-one-thing sequence is what we used for the newly walking baby.

FAQ: keeping a baby sleep log

How long should you keep a sleep log?

Two weeks. One week is too short to separate a bad run from a pattern; a month is long enough that most parents abandon it and that the child’s needs shift underneath the data.

What should you write in a baby sleep log?

Morning wake time, each nap’s start and end, when the bedtime routine began, when the child was actually asleep, the clock time of each night waking, and what you did — in three words or fewer.

Should I use a sleep tracking app instead of paper?

Either works, but apps tempt you to record far more than the six things that matter, and the extra fields are what make a log unreadable. Paper by the bed is the version people actually finish.

Does keeping a sleep log fix anything by itself?

Occasionally, because it exposes a bedtime that had drifted. Mostly it tells you which single change is worth making, which is a much better position than guessing.

Is this the same as Montessori observation?

It’s the same discipline applied to a different setting: record what happened, not why you think it happened, and read the record before you act. In a classroom that means watching a child work without intervening; at home, at 2am, it means writing down a time and nothing else.