When Sleep Training Isn't Working: What I Check First
When sleep training isn’t working, the cause is almost never the method itself. In consults, the pattern I see is one of five things: an unresolved medical question, execution that isn’t as consistent as the household believes, a schedule that doesn’t match the baby’s current age, a method the adults can’t actually run at 2am, or an environment and a set of “exits” that quietly undo the work each night. I check those in that order — and I do not let anyone switch programs until I’ve been through them, because switching on night four is the single most expensive move in this field.
I’m a working sleep consultant, which means troubleshooting other people’s stalled plans is most of my week. Here’s the order I use, unedited.
Is it failing, or is it only night three?
Before troubleshooting anything, establish that there’s something to troubleshoot. Almost every structured plan gets worse before it gets better: a hard night one, a deceptively good night two, and a night three that makes parents want to burn the whole thing down. That is the normal curve, not a failure signal.
The honest test is a week, judged on trend rather than on last night. Are protest minutes shrinking across the week, even messily? Are night wakings fewer, shorter, or easier to resolve? Is bedtime itself getting quicker? If any of those lines is bending the right way, you don’t have a broken plan — you have an unfinished one. If a genuinely consistent ten to fourteen days moved nothing at all, that’s a real signal. Now troubleshoot.
The order I check things in
1. The medical door, first and every time
If a plan isn’t budging, the first thing I want ruled out is discomfort. Reflux beyond ordinary spit-up, noisy or laboured breathing, snoring, suspected apnea, ear pain, eczema bad enough to wake a child, feeding difficulty, weight-gain concerns, anything that started with a new medication — none of these are behavioural problems, and none of them yield to a check-in schedule. Crying that comes from pain is not protest, and no method should be layered over an unexamined symptom.
That is a pediatrician conversation, full stop. If you have even a low-grade suspicion in this category, book the appointment before you spend another week on a plan. The full not-yet list lives in my checklist of when not to sleep train, and a good share of the stalled plans I see should never have been started in the first place.
2. Consistency, honestly measured
Every parent tells me they were consistent. Most of them were, most of the time. “Most of the time” is what stalls plans.
The three leaks I look for: variation between caregivers (one parent doing timed checks while the other picks up), variation between bedtime and the 3am wake (a firm bedtime followed by a rescue teaches that the middle of the night runs on different rules), and variation between nights (four nights of the plan, one night of “we were wrecked, she came into our bed”). Intermittent rescue is the most powerful teaching signal in the house, and it teaches the opposite of the plan.
The fix isn’t more willpower. It’s a smaller plan the household can run identically for ten nights — including at 3am, including when only one adult is home.
3. The schedule, before the method
A method asks a baby to fall asleep. A schedule decides whether they can. Wake windows too short means too little sleep pressure — you’re asking for something the biology can’t deliver; much too long gives you an overtired, wired child who fights harder and wakes more. Both look identical from the hallway, and both get misread as “the method isn’t working.”
Age-appropriate day sleep is the lever I pull most often before touching anyone’s method, and it is free: total daily sleep in the right neighbourhood, day sleep not stealing from night, a bedtime that matches the last window rather than the clock on the wall, and a wake-up time you hold steady. If overtiredness is the suspect, Betteroo’s explainer on the overtired baby covers the paradox of a tired child sleeping worse better than most.
Also check the calendar. A nap transition landing mid-plan, or the first week of a developmental leap, will make a good method look useless.
4. Method mismatch, not method failure
Some pairings are simply wrong. A no-cry approach in a household that can’t tolerate a three-to-six-week timeline gets abandoned at week two and recorded as a failure. Parental-presence methods backfire with the babies who find a visible parent maddening rather than soothing — if your child escalates when you enter the room, the chair is the wrong tool and persistence won’t fix it. And a strict extinction protocol in a household where one adult can’t stand it produces the intermittent rescue described above.
Mismatch is worth a change; disappointment isn’t. The difference is whether you can point to a mechanism — timeline you can’t sustain, escalation on parental presence, a plan neither adult will defend at 2am. My comparison of the main methods is organised around exactly that question, and switching once, deliberately, after a diagnosis is entirely reasonable. Switching monthly is how families spend a year training.
5. The environment and the exits
Last, the cheap physical stuff, because it’s cheap. Light leaking in at 5am, a sound machine too quiet to mask a sibling or a street, a room warm enough to be uncomfortable. For older babies and toddlers, add the exits: a crib they can climb, a door that opens, a caregiver who stays for “one more song.” Every exit is a slot machine, and slot machines are the hardest thing in psychology to switch off.
When to stop and rebuild
Sometimes the answer is to stop. If the baby is ill, if a medical question is open, if the household is in genuine crisis, or if two weeks of honest effort produced nothing — pause deliberately rather than grinding on. A clean stop is not failure; it’s the difference between one hard week and two exhausted months. Rebuild from the groundwork in the free sleep stack: routine, light, schedule, environment. Plenty of plans that failed in month one succeed in month three simply because the schedule finally fits.
And if you’ve run all five checks, fixed what you found, and it’s still broken, that is the case where a human — a consultant, and not necessarily an expensive one — earns their fee. Diagnostic judgment on a stuck case is the actual product my industry sells. Buying it before the five checks is buying it too early.
What I’d tell a client
Don’t change the method. Change one variable, then wait five nights. Nine times out of ten the variable is the schedule or the consistency, not the program you bought — and both are free to fix. Judge on the week’s trend rather than on last night, keep a plain log so you’re arguing with data instead of memory, and take anything that smells medical to your pediatrician before you take it to me.
FAQ: when sleep training isn’t working
How long should sleep training take before I decide it failed?
Give a consistently executed method ten to fourteen nights before you judge it, and judge the trend rather than the worst night. Most plans show some movement — shorter protests, fewer wakings, faster bedtimes — within a week. No movement at all after two honest weeks is a real signal to troubleshoot.
Why did sleep training work and then stop working?
Usually an age-driven schedule change (a nap transition, or wake windows that outgrew the plan), an illness or teething stretch that reintroduced night help, or travel. Regressions and developmental leaps also interrupt a settled pattern temporarily. Re-establish the schedule first; you rarely need to retrain from scratch.
Should I switch methods if sleep training isn’t working?
Only after you can name the mismatch — an unsustainable timeline, escalation when you enter the room, or a protocol one adult won’t run. Switching without a diagnosis restarts the clock and teaches the child that protest changes the plan.
Can sleep training fail because something medical is going on?
Yes, and it’s the possibility I check first. Reflux, ear infections, breathing issues, allergies and pain don’t respond to behavioural plans, and a child in discomfort will out-persist any protocol. Rule that out with your pediatrician before you assume the method is at fault.
