A baby sleep training schedule is not the method, and confusing the two is the most common reason a first attempt falls apart. The method is what happens in the ninety seconds after you put her down. The schedule is everything else — what time she woke, how long she was awake before the last nap, how long that nap ran, and whether the bedtime you picked was the one her day genuinely produced or one you chose in advance because the internet said seven o'clock. You can execute a method flawlessly against a badly built day and get nothing back, and many of the families who say sleep training did not work for them are describing exactly that.
So here is the answer before the detail. For roughly a week before you begin, hold a morning wake time inside a thirty-minute window and change nothing else. Build the day forward from that wake time using age-typical awake stretches, and let bedtime fall out of the arithmetic instead of assigning it. Then, once the nights start, freeze the day: same wake time, same naps, same bedtime derived the same way, for at least seven nights, no matter how the naps behave in the middle of the week. One variable at a time, and make the night the only one that moves.
This page is about that day. It is deliberately not another walkthrough of check-in intervals — the methods themselves, compared neutrally and including the option of not doing any of them, are on how to sleep train a baby. What you get here is the logistics: the pre-training week, the anchors worth defending, an hour-by-hour example of a training week, what to do when the naps collapse partway through, and a diagnostic for the question that actually matters at two in the morning — is this a schedule problem or a settling problem? From the hallway they look identical. Their fixes are opposite.
One thing to say plainly at the top. None of this obliges you to sleep train. The American Academy of Pediatrics does not require any behavioural sleep intervention, and neither does this site. A well-built day is worth having on its own terms, and the last section of this page is about using the schedule half if you have decided against the training half entirely.
Need the clock times for your baby?
The calculator builds the whole day from one wake-up.
The week before night one
Nothing that happens in a training week is interpretable if the days underneath it are still moving. That is the entire argument for a run-up, and it is worth about seven days.
Three things happen in that week. You pick a morning wake time and hold it inside a thirty-minute window, waking her if she runs past the top of it. You build the day forward from that anchor and let bedtime land where the arithmetic puts it. And you change nothing else — not the room, not the sleep sack, not the feeding pattern, not the order of the routine. If you want the mechanics of building a day from scratch, how to create a baby sleep schedule covers it. The point here is narrower: a baby sleep training schedule has to be boring and repeatable before the night can tell you anything at all.
The reason is attribution. Change the day and the night in the same week and, if something improves, you will not know which one did it — and far worse, if nothing improves you will not know what to adjust. Sleep training costs something real in everybody's distress. Spending that on an uncontrolled experiment wastes a week you do not get back.
There is also a list of weeks not to pick. An active illness, the days around an immunisation, teething that is visibly bothering her, a house move, travel across time zones, a nap transition already underway, or one parent about to leave for a work trip — any of these makes the result unreadable and the week harder than it has to be. Wait. Nothing is closing.
Three anchors, and only one of them is a clock time
A training week needs three fixed points, and most people fix the wrong ones.
The first is the morning wake time. This is a real clock time, and it is the only one you actually choose. Everything downstream is arithmetic from it, which means letting it float by an hour on a Sunday quietly moves the whole day — including bedtime, including when the first night waking lands, including whether night four is comparable to night three.
The second is the last awake stretch of the day: the gap between the end of the final nap and lights out. It is not a clock time, it is a duration, and it is the single most load-bearing number in a training week.
The third is bedtime, and this is where it usually goes wrong, because bedtime is an output rather than an input. Cleveland Clinic — one of the very few major institutions to publish anything quantitative about infant awake time — declines to name a bedtime and prescribes a method instead: start from when the child needs to wake, and count backwards. The seven-to-eight-o'clock rule repeated across the internet has no authoritative source behind it. The only published clock range I can find is Raising Children Network's observation that most babies are ready for bed between 6pm and 10pm — honest, and almost useless as a prescription, which is rather the point.
Why this matters specifically for training rather than generally: a baby put down forty minutes before she is tired will protest, at length, and that protest is indistinguishable from protest at the method. You will conclude the method failed. It never got a turn.
Wake windows and night settling are the same problem in two hats
First, the caveat that accompanies any wake-window number on this site. Wake windows are not an established medical construct and no consensus body publishes them. Among major institutions only Cleveland Clinic publishes an explicit chart; Raising Children Network frames the same idea qualitatively, as the time before tired signs appear; most hospital sources decline to quantify awake time at all. The minute-level tables dominating search results are commercial products. Treat the ranges below as commonly-published guidance held deliberately wide — from six months onward Cleveland Clinic and Raising Children materially disagree, and averaging them into a tidier midpoint would invent a precision nobody has.
The interaction itself is real, and it is why a schedule page belongs beside a training page. Too little awake time before bed and there is not enough sleep pressure to fall asleep quickly; the crying that follows is frustration rather than distress at being left, and it runs long because the biological push toward sleep is not there yet. Too much, and you get the opposite failure — the baby parents describe as wired, arching away from the crib, fighting sleep hardest exactly when she most obviously needs it.
The uncomfortable part is that these two opposite causes produce a similar-looking bedtime. They diverge in what happens afterwards, which is how the table separates them. The two right-hand columns are pattern recognition drawn from what parents commonly describe, not measured findings — no source in this site's register publishes a symptom-to-cause mapping, and I am not going to pretend one does.
| Age | Commonly published awake stretch | If the last stretch is too short | If it is too long |
|---|---|---|---|
| 4–6 months | 1.5–2.75 hours | Long, unfrantic protest; still awake 45 minutes after lights out | Frantic crying and arching, then a very early first waking |
| 6–8 months | 2–4 hours (sources disagree; held wide) | A bedtime battle with a baby who is plainly not tired | Falls asleep fast, then wakes 45–60 minutes later |
| 8–10 months | 2–4.5 hours (the widest disagreement in the whole range) | Bedtime drifts later on its own, night after night | Waking before 5am and refusing to resettle |
| 10–12 months | 2.5–4.5 hours | Nap refusal that looks like a nap drop and is not one | Bedtime protest plus a fragmented first half of the night |
A training week, hour by hour
What follows is one concrete day, worked from a single decision, for a baby around six months old still taking three naps. It is arithmetic, not a prescription. The point is to show how the pieces connect — change the wake time and every row below moves with it.
Say you choose 06:45 as the morning anchor. At six to eight months the commonly published awake stretches run roughly two to four hours, and the convention is shortest stretch first, longest last, because sleep pressure is lowest right after a night of sleep and highest by evening. That gradient is a convention, not a published finding, and it is worth knowing which is which. Total daytime sleep here is commonly put at around two and a half to four hours across all the naps.
So the day builds itself. Awake at 06:45. First nap down around 08:45 after a two-hour stretch, up by 10:00. A slightly longer stretch, second nap around 12:30, up by 14:00. Longer again, a short third nap at 16:15, up by 16:45. That leaves a final awake stretch of about two and a half hours, which puts lights out near 19:15 — inside Raising Children's 6pm-to-10pm range, without anybody having decided on seven o'clock in advance.
During a training week you run this same day whether or not the night went well. The routine starts at a fixed point relative to lights out rather than at a fixed clock time, so it travels with the day instead of pinning it in place. That single detail is what makes night three comparable to night one.
| Clock | What happens | Where the number comes from |
|---|---|---|
| 06:45 | Wake and open the curtains, including after a bad night | The only time you choose; everything else is derived from it |
| 08:45–10:00 | Nap one | Shortest awake stretch placed first — convention, not a published finding |
| 12:30–14:00 | Nap two | The longest nap of the day, sitting mid-day |
| 16:15–16:45 | Nap three, short | Exists to stop the final awake stretch running too long |
| 18:45 | Routine begins: bath, sleep sack, feed, one book, lights down | Fixed relative to lights out, not to the clock |
| 19:15 | Down awake — this is where the method starts | Output of the day's arithmetic, inside the 6–10pm published range |
| Overnight | Whatever your chosen method prescribes; tomorrow's day does not change | Holding the day constant is what makes night-to-night comparison mean anything |
Leave the naps alone for the first week
Naps often get worse in the first days of night training. Parents report it consistently. No source in this site's register measures how often or for how long, so hold it as something to expect rather than something known — but do expect it, because families who are not warned read it as the whole project backfiring and stop on night four.
The instruction is the same either way: do not start fixing naps in the same week. You have one variable in play. Adding a second means neither result is readable. And nap training is materially harder than night training on its own terms — the drive to sleep is weaker in the daytime, the room is brighter, the house is louder, and a failed nap has a hard stop because the day simply carries on regardless.
Absorb the damage at bedtime instead. If the naps came up short, the final awake stretch comes down too, and bedtime moves earlier to compensate. That is not a defeat or a cheat; it is the arithmetic doing its job. A day that produced ninety minutes of daytime sleep instead of the usual hundred and sixty-five has to repay that time somewhere, and the only place available is the front of the night. Lights out at 18:15 on a wrecked day sits well inside the published bedtime range and usually beats holding 19:15 out of stubbornness.
The one thing not to do is let her sleep in the following morning to make up for it. That moves the anchor, which moves the entire day, which means night two is being run against a different schedule than night one and you have lost the comparison.
Schedule problem or settling problem?
This is the question the whole page exists for, and there is a reasonably reliable way to answer it.
The shorthand: schedule problems show up in the timing of the trouble, settling problems in its shape. A baby whose day is wrong has trouble at a predictable point — always at bedtime, or always at the same hour, or always at 04:50 — because the cause is a clock, and clocks are consistent. A baby who has not learned to return to sleep unaided has trouble wherever a sleep cycle happens to end, scattered rather than fixed, and the giveaway is that she settles almost instantly when you reinstate whatever she fell asleep with, then wakes again one cycle later.
There is a cheap test for the second one. If a waking ends within about a minute of you picking her up, feeding her, or replacing a pacifier, then recurs an hour or two later, that is a sleep-onset association pattern. No schedule adjustment resolves it, because the schedule is not causing it — that is precisely what sleep training addresses. If instead a waking takes a long time to end whatever you do, or lands at the same clock time nightly, or is reliably the first waking at forty-five to sixty minutes after bedtime, look at the day first.
The order matters. Fix the schedule first, then decide about training — partly because a fixed schedule sometimes resolves the problem outright and saves you the whole exercise, and partly because training against a broken day is a very expensive way to discover that the day was broken.
| What you are seeing | Points to the schedule | Points to settling | Try first |
|---|---|---|---|
| Takes 40+ minutes to fall asleep at bedtime, not especially distressed | Yes — usually too little awake time before bed | Rarely | Extend the last awake stretch by 15–20 minutes for three nights |
| Frantic at bedtime, crashes hard, wakes 45 minutes later | Yes — usually too much awake time | Sometimes both at once | Move bedtime 20 minutes earlier and hold the wake time |
| Wakes two to four times, resettles instantly when fed or held | No | Yes — the classic association pattern | This is what training addresses; the schedule will not touch it |
| Wakes at the same clock time every single night | Yes — check total daytime sleep and bedtime | Unlikely | Recalculate the day before changing anything at night |
| Up for the day between 04:30 and 05:15 | Yes — commonly bedtime too late, or too much daytime sleep | Sometimes | Earlier bedtime first. It is counter-intuitive and it is usually right |
| Naps are fine, nights are not | Unlikely | Yes | Look hard at what she falls asleep with at bedtime |
| Nothing is consistent from night to night | Yes — the day is probably still moving | Cannot tell yet | Go back to the pre-training week and stabilise before judging anything |
The feed you are keeping
Night feeds are the one part of a training schedule that is not yours to settle from a chart. Whether a baby still needs calories overnight depends on age, weight, growth trajectory, feeding method and medical history, and it belongs to your pediatrician. This site does not answer it, and nobody selling you a program should be answering it either.
What a schedule can do is stop a retained feed behaving like a variable. If you and your pediatrician agree one overnight feed stays, give it a fixed position rather than feeding on demand across the night. Pick a time. Feed at that time whether or not she woke for it — some families wake her deliberately, worth raising at the same appointment — and handle every other waking by whichever method you chose. The night now holds one planned event and one experiment, instead of an unbounded series of judgement calls made at 03:00, which is when judgement is worst.
The failure mode is migration. A feed that starts at 02:00 and drifts to 01:15, then 00:40, then twice a night, has stopped being a fixed point and become the thing you are troubleshooting. Write the clock time down — a note on your phone is enough. Memory at three in the morning is not evidence, and two adults will remember the same night differently.
Daytime intake is the other half of this and it changes faster than people expect once solids arrive. If night feeds are increasing rather than decreasing, that is a conversation to have, not a schedule to tighten.
If you have decided not to sleep train
Everything above still applies, minus the nights.
A stable morning anchor, a day built forward from it, and a bedtime derived rather than declared are not sleep-training techniques. Strip the nights out and a baby sleep training schedule is simply a schedule — the part of this picture with the least controversy attached. Families who have decided against any formal intervention — because it conflicts with how they want to parent, because they tried it and found it intolerable, because close night-time contact is the assumption in their family or culture, or because their baby's sleep is honestly not the largest problem in their life this month — get exactly the same benefit from the day-side work as anyone else.
Without hedging: no authoritative body requires sleep training. The American Academy of Pediatrics publishes safe-sleep guidance and general infant sleep information without mandating any behavioural protocol, and the literature does not show that infants who were never sleep trained fare worse. Pennestri and colleagues examined this directly in Pediatrics in 2018 and found no association between uninterrupted infant sleep and infant development on their measures. The association they did find was with maternal mood — which cuts both ways. It is a reason not to feel obliged, and equally a reason to treat a depleted parent as a real cost worth solving.
If you want the longer version of that argument, including what the evidence can and cannot support in either direction, it is on how to sleep train a baby, and the homepage collects the rest of the schedule tools. Build the day. Decide about the nights separately, or not at all.
Questions parents ask
How stable does the schedule need to be before we start?
About a week of days that look alike is a reasonable bar. Concretely: a morning wake time that has not varied by more than thirty minutes, naps that start within roughly half an hour of the same points, and a bedtime you can predict the night before. It need not be perfect; no baby produces identical days. It needs to be stable enough that if night four goes badly you can rule the day out as the explanation, rather than spending the week guessing at two variables at once. If you are still moving nap times around to see what works, you are not ready to start the nights.
Should bedtime be earlier during a sleep training week?
Often, yes — but as an output rather than a decision. If naps came up short, the day produced less daytime sleep, and the arithmetic pushes bedtime earlier on its own. Following that is not cheating. An overtired baby at bedtime protests harder and longer for reasons that have nothing to do with the method, which makes the week harder and the result less readable. Moving lights out twenty to forty minutes earlier on a bad-nap day sits well inside the published bedtime range and usually buys a calmer night one.
Do I need to fix naps before starting on nights?
No, and trying to do both at once is one of the reliable ways to get an uninterpretable week. Naps need to be consistent enough that the day is predictable — starting near the same times, running near the same lengths — but they do not need to be good. Nights are the easier target: the drive to sleep is strongest, the room is dark, and there is no hard stop the way there is when a nap fails at two in the afternoon. Fix nights, hold naps steady, revisit naps a fortnight later.
What time should night one actually start?
At the bedtime your day produced, not at a time chosen for its convenience. Work the day forward from the morning wake time, subtract the final awake stretch from the end of the last nap, and that is your lights out. Then start the method there. The temptation is to start early so more of the evening is left to survive it, but putting her down before she is tired guarantees a long protest that tells you nothing about the method.
Her naps fell apart on night three. Do we stop?
Almost certainly not. Naps getting worse in the first days is commonly reported, though no register-quality source measures how often or for how long, so treat it as a thing to expect rather than a documented pattern. The response is to absorb it at bedtime — shorter final awake stretch, earlier lights out — and hold everything else. Stopping on night three is the single most common way a training attempt is abandoned at exactly the point where families most often describe things beginning to turn.
Should I wake her in the morning during a training week?
Yes, if she runs past the top of your thirty-minute window. It feels cruel after a bad night and it is the thing that keeps the week comparable. Letting her sleep in shifts the entire day — later naps, later last awake stretch, later bedtime — which means tonight is being run against a different schedule than last night, and you lose the ability to tell whether anything changed. Open the curtains, get light into the room, and start the day. The compensation goes at the front of the night instead.
Bedtime keeps drifting later. What do I change?
Look at the morning first, because a drifting bedtime almost always traces back to a drifting wake time. If the wake time is genuinely fixed, the next suspect is the final nap running too late or too long, which pushes the last awake stretch and everything after it. Cap the last nap by clock time rather than letting it end on its own. If both are stable and bedtime still creeps, the awake stretches may be too short for her age.
Can we sleep train while she is dropping a nap?
It is worth waiting. A nap transition is a period of genuine instability — daytime sleep totals move, the last awake stretch stretches, bedtime swings, and the days are inconsistent by definition for a few weeks. Running a training week inside that produces a result you cannot attribute to anything. Transitions take weeks rather than days, so waiting for the far side is not an indefinite delay. If one has run a month with no sign of settling, the schedule pages are the better starting point.
Is it better to start on a Friday?
For most households, yes, though the reasoning is about the adults rather than the baby. The first two or three nights are usually the hardest, and having them land when nobody has to be functional at nine the next morning removes a real pressure that otherwise pushes families to abandon the plan early. Two other things matter more than the day of the week: that both adults are home and running the same protocol, and that the week ahead is clear of travel, immunisations and visitors.
How long do I hold the same schedule before judging it?
Seven nights, unmodified, is the working minimum, and the whole page is built to make that possible. Judging earlier is not really judging — it is reacting to the two nights that are always hardest. Seventy-two hours is the shortest interval at which a schedule change alone tells you anything, and a night-time behavioural change generally needs longer. Keep a one-line log nightly — down time, how long she took to fall asleep, when she woke. That log, not your recollection, is what you review on night eight.
What if the schedule is right and it still is not working?
Then it is worth separating the three things that sit outside any schedule. First, medical: reflux, an ear infection, an undiagnosed feeding issue and obstructed breathing all present as unexplained night waking, and none of them respond to timing. Second, the method itself may be a poor fit and worth swapping rather than intensifying. Third, the adults may not be running the same protocol, which is more common than anyone admits. Take the log to your pediatrician before you take it to another schedule.
Sources
Know a tired parent who needs this?
Send it their way — it is free and always will be.
Keep reading
Building a full day from scratch? Start with the baby sleep schedule guide and calculator on the homepage.
Written by
Sukie Gao
Mom of a daughter, and the person behind every tool on this site.
Sukie Gao has a daughter, and when her daughter was a newborn she was exhausted and completely out of her depth. She spent months piecing together sleep advice from scattered corners of the internet, and built My Baby Sleep Schedule so the next tired mother would not have to search as hard. She writes as a parent from lived experience, not as a clinician.
Last reviewed and updated . Sukie is a parent, not a clinician — check anything that matters with your pediatrician.