You’ve read the books. You’ve watched the videos. You’ve practiced the “drowsy but awake” handoff like it’s a magic trick. Lower the baby in. Withdraw your hands slowly. Tip-toe out.

And every single time, before your foot crosses the threshold, the eyes pop open. The cry starts. You’re back at the crib in fourteen seconds.

If this is night ten of trying drowsy-but-awake and you’re starting to wonder if your baby is broken — they’re not. The advice itself is fine for some babies and a disaster for others, and it’s almost never explained which is which.

Here’s what’s actually going on, and four alternatives that work better for the babies drowsy-but-awake can’t reach.

What “drowsy but awake” is supposed to do

The theory is straightforward. Babies (like adults) learn to associate where and how they fall asleep with where and how they expect to be when they wake up. If your baby falls asleep in your arms with the bottle, when they cycle into light sleep at the 45-minute seam, they’ll look for arms and bottle. Not finding them, they fully wake.

If they fall asleep in the crib, alone, on the mattress — they’ll cycle through that seam without needing you.

So the prescription is: get them into the crib just before they actually fall asleep, so the last few seconds of consciousness happen in the place they’ll wake up. Sleep onset cues match the sleep environment. They self-regulate through cycles. Eight straight hours, in theory.

This works for some babies. Not most, in our experience and in every parent group we’ve been part of. The advice has never actually been studied in isolation as its own intervention, so anyone giving you a percentage is making it up. The honest answer is: maybe a third.

Why drowsy-but-awake fails for some babies

There are five reasons, in roughly the order they show up:

1. Your baby is under 4 months

Before about 16 weeks, sleep architecture is fundamentally different. Newborns spend roughly half their time in active sleep — the developmental precursor to adult REM — and the deeper NREM stages haven’t fully developed yet. The cycles are also short, around 40 to 50 minutes, and the lighter stages dominate.

Drowsy-but-awake at 8 weeks is asking a baby to do something their nervous system hasn’t built the hardware for yet. This isn’t a failure. It’s a developmental mismatch. The advice is calibrated for 4 to 6 month olds and older, even when the books don’t say so.

2. The “drowsy” window is microscopic

The window — eyes glazed, breathing slowing, body heavy but not floppy — is genuinely brief. There’s no clinical measurement of how long it lasts; it varies. What you find out, every time, is that you missed it by the cry that starts ninety seconds after you sit down on the couch.

Too awake at transfer: the baby is alert enough to register the change in environment and protests immediately.

Too asleep at transfer: you’ve created the same association you were trying to avoid (baby learns “I fall asleep in arms, I wake up in crib,” which is its own confusion).

3. Your baby is wired to register every transition

Some babies clock everything. The light shift, the temperature change between your chest and the mattress, the sound of your shirt brushing the crib rail. For these babies, the act of being lowered into a crib is the wake-up event, regardless of how drowsy they were on the way down.

It’s not a discipline thing. It’s how they’re wired. The baby who tracks every dog walking past the stroller is the same baby who clocks the temperature drop in a 0.3-second handoff. Sleep researchers call this perceptual sensitivity, but you don’t need the term to recognize it.

4. There’s something else going on

Reflux, an ear infection, the 3-to-4 month sleep change. If any of these are active, no transfer technique will work consistently because the baby isn’t ready for sustained sleep regardless of how they got there.

You can’t sleep-train through teething. You can’t sleep-train through a real regression. Rule these out first or you’ll spend a week blaming a technique that never had a chance.

(If you suspect reflux, that’s a pediatrician conversation, not a sleep blog conversation. Some of these need diagnosing before any sleep approach makes sense.)

5. The sleep environment isn’t doing its job

Drowsy-but-awake leans heavily on the environment to take over the soothing role you just dropped. If the room is too bright, too quiet, too warm, or too unfamiliar, the baby has nothing to anchor to once you’re gone.

This one is the easiest to fix and the most often overlooked.

Four alternatives that work when drowsy-but-awake doesn’t

If you’ve ruled out reasons 1 through 5 and the technique still isn’t landing, you have options. Each one trades off a different way.

Option 1: Fully asleep transfer, then re-train at 5 to 6 months

In the early months, just get them all the way to sleep, do the careful transfer, and accept that they’ll wake at the cycle seam sometimes. At around 5 to 6 months, when sleep architecture has actually matured and they’re physically ready for self-regulation, you re-train.

Easier now, harder later. You’ll have to undo the rocked-to-sleep association eventually, and that conversation with your future self is real. But a 6-month-old can learn things an 8-week-old physiologically cannot, and if you have a newborn, or you’re running on three hours and you just need to sleep — this is the right answer right now.

Option 2: The chair method (sometimes called camping out)

You stay in the room. The first few nights, hand on the chest, shushing, full involvement. Each subsequent night, you do less — hand off the chest, then sitting next to the crib, then by the door, then outside the door. You’re fading your own presence on a schedule, not the baby’s sleep onset.

Takes 1 to 3 weeks. Doesn’t usually involve sustained crying. Best for sensitive babies (the ones who fail drowsy-but-awake for reason 3) and parents who can’t tolerate cry-it-out methods. The catch is that you have to actually commit — bailing out on night 4 because it’s hard means you reset the clock, and the baby has just learned that escalation works.

Option 3: Pick-up-put-down

When the baby cries after being put down, you pick them up and calm them until they’re calm but awake, then put them back down. Repeat as many times as needed. It can take 30 to 90 minutes a night for the first week.

There’s less research on this one than on Ferber-style approaches, but it’s a common middle ground for parents who want active involvement without sustained crying. The honest caveat: some babies get more activated by the picking-up than soothed, and in those cases this method makes things worse before it makes them better. If you’re three nights in and the picking-up is winding the baby up rather than down, switch.

Best for babies 4 to 6 months and up — too young for this and you’re doing Option 1 anyway.

Option 4: Modified extinction with check-ins (the Ferber approach)

Put baby down awake, leave the room, return at increasing intervals (3 minutes, 5 minutes, 10 minutes, then every 10 from there) to briefly reassure without picking up. Keep going until they fall asleep.

The research on extinction-based methods (this and the unmodified version) is the strongest of any sleep training approach. Improvement typically shows up by night 3 or 4; full resolution within 1 to 2 weeks.

It’s also the method most likely to gut you. The first night you sit on the stairs with a baby monitor, count the seconds to the next check-in, and feel like a monster — every parent who’s done it knows that exact stairwell. The research says outcomes are good. The lived experience is hard, and pretending otherwise doesn’t help anybody.

How to pick which alternative fits your baby

Three quick questions:

How old is your baby? Under 4 months — Option 1 only. The others aren’t developmentally appropriate yet. Over 6 months — all four are on the table.

How does your baby respond to extended crying? Some seem to escalate rather than settle — they cry harder the longer it goes, sometimes to the point of vomiting. For those babies, Options 2 or 3 are likely the better fit. Other babies de-escalate within 10 to 20 minutes, and Option 4 is faster and often kinder over the full timeline.

What’s your honest tolerance as a parent? If you can’t hold the line for Option 4 at 1am on night 3, don’t start Option 4. Inconsistency tends to backfire — bailing mid-method risks reinforcing the crying you were trying to extinguish, which is the opposite of what you wanted to teach. Be honest with yourself about which method you can actually live through.

The 7-night (or 14-night) experiment

Whatever you pick, hold it for at least seven nights. Sleep is too noisy night-to-night to evaluate any method on a single trial. The pattern is in the trend, not the day.

For seven nights, write down the same five things: when bedtime started, when the baby actually went down, how long that took, how many times they woke up, and the total sleep. That’s it. Not because the data is magic, but because at 4am on night five you will not remember whether Tuesday was better or worse than Monday, and the only way to know if a method is working is to compare nights you can’t trust your memory on.

Extinction-based methods (Option 4) should show movement in the trend by night 7 — faster sleep onset, fewer wakeups, or longer stretches between them. Gentler methods (Options 2 and 3) often need 10 to 14 nights. Don’t bail early on a gentle method.

If you don’t see any movement after the relevant window, the method probably isn’t the right fit and you can move on without guilt.

Where Robin Baby fits

This kind of tracking is the thing that falls apart at 2am. You’re not opening a spreadsheet. You can barely find the lamp. Robin Baby’s voice logging is one tap and a sentence: “she went down at 7:42, slept 11 to 4.” Done. The 14-day trend view shows you whether the method is moving the needle before you give up on it on a bad night.

Voice logging is part of Robin’s Pro tier (free 7-day trial). The free tier — manual logging, the timeline, the sleep forecast — covers most of what you need to actually run the experiment.

The honest bottom line

Drowsy-but-awake is one technique among several, not the gospel it’s sometimes presented as. If it’s failing you, the failure isn’t in your execution. It’s almost always a developmental mismatch, a temperament fit, or an environmental issue. Once you know which, the alternatives are real and they work.

The hardest part isn’t picking the method. It’s giving any one method long enough to actually evaluate it. Seven nights, one variable at a time, written down.

You’re not doing it wrong. You just got handed advice that wasn’t built for the baby in front of you.

'Drowsy but awake' doesn't work for my baby — what to try instead