If bedtime takes an hour of protest every night, the most common cause is not a baby who resists sleep. It is a bedtime that arrives before your baby is tired enough to use it. Bedtime fading fixes that by doing the thing every instinct says not to: putting bedtime later, on purpose, before walking it back.
What it is
You find the time your baby actually falls asleep, set bedtime to that time, and let them fall asleep quickly for a few nights. Once settling is fast and reliable, you move bedtime earlier in small steps until it lands where you want it. The mechanism is sleep pressure: a baby put down when genuinely tired falls asleep, and the association being built is bed means sleep rather than bed means an hour of negotiation.
What the evidence shows
A randomised trial assigned 43 infants aged 6 to 16 months to graduated extinction, bedtime fading, or a sleep-education control. Both interventions produced large decreases in sleep latency — the time it takes to fall asleep — compared with the control group.[1]
Large drop
The trial measured infant stress directly through salivary cortisol, and bedtime fading showed small-to-moderate declines compared with controls — stress went down. Twelve months later there were no differences between the groups in emotional or behavioural problems, and no differences in attachment measured with the strange situation procedure.[1] The AASM review of 52 studies also lists bedtime fading and positive routines among the specific therapies the evidence supports.[2]
One honest limitation. In the same trial, graduated extinction produced large decreases in night wakings and time awake after falling asleep, while bedtime fading's clear win was on how fast the baby fell asleep at the start of the night.[1] If your problem is a baby who goes down easily but is up four times a night, this is probably not your method.
Finding the real bedtime
This is the step people skip, and skipping it is why the method fails. You are not guessing at a good bedtime — you are measuring the one your baby already has.
Setting the starting point
Do this before changing anything. A week of honest logging is worth more than any published schedule, because the number you need is specific to your baby this month.
- 1
Log actual sleep onset for five to seven nights
Not when you started the routine and not when you left the room — the time your baby was genuinely asleep.
- 2
Take the average, then round to the later side
If your baby has been falling asleep between 8:15 and 8:50, your starting bedtime is about 8:45, not 7:30.
- 3
Move bedtime to that time
This will feel wrong. You are temporarily giving up the early bedtime you wanted in order to get it back properly.
- 4
Keep the full routine, just later
Same sequence, same length, same room. Only the clock time changes, so bedtime still signals sleep.
- 5
Hold until settling takes under about 15 minutes
Usually two or three nights. This is the point at which bed reliably means sleep, and it is what the rest of the method is built on.
Walking it back
Once your baby is falling asleep quickly and consistently, start moving bedtime earlier. Slowly enough that settling stays fast at every step.
| Nights | Bedtime | Move on when |
|---|---|---|
| 1–3 | 20:45 | asleep within ~15 minutes, three nights running |
| 4–6 | 20:30 | same test |
| 7–9 | 20:15 | same test |
| 10–12 | 20:00 | same test |
| 13–15 | 19:45 | stop here if this is your target |
Fifteen minutes every three nights is the usual pace. If settling stretches back out at any step, you have moved too fast — go back to the previous time for a few nights and then try a smaller step. The end point is wherever settling stays quick, which may be earlier or later than the bedtime you originally had in mind.
Why an early bedtime backfires
The instinct when a baby is not sleeping well is to put them to bed earlier, and it is often exactly wrong. A baby who is not yet tired will lie there, get frustrated, need company, and learn that bed is where you go to be bored. Meanwhile the last stretch of the day is doing the actual work — which is why the last wake window matters more than the others and should usually be the longest, as wake window stacking sets out.
This is also why bedtime fading pairs badly with a late or long final nap. If the nap is stealing the sleep pressure that bedtime needs, cap or move the nap first and then measure again, or you will be fading a bedtime against a moving target.
Who it suits
- Households where crying is the reason no method has been attempted at all.
- Babies who take 45 minutes or more to fall asleep but then sleep reasonably well.
- Bedtimes that have drifted early over months, often after a nap transition.
- Anyone who has abandoned an interval method partway. An abandoned method teaches nothing, and this one is far easier to finish.
- Toddlers, where it tends to work well and interval methods get harder as climbing out of a cot becomes an option.
When it is not enough
If bedtime settles quickly but your baby is still waking repeatedly in the night, bedtime fading has done its job and the night wakings need a different tool — that is where the Ferber method had its clearest advantage in the trial, and the chair method is the option that keeps a parent in the room while addressing them.[5] The methods compared sets out which problem each one actually solves.
And keep the wider frame in view. Behavioural sleep techniques buy short- to medium-term improvement rather than permanent change: in one trial the control group had caught up by four months,[6] and a five-year follow-up found no differences from controls on any outcome measured.[4] Choose the method your household can sustain, and judge it on the 24-hour total rather than on any one bedtime.[7]
