TL;DR. Most circulating advice on insomnia runs on the logic of “go to bed earlier, sleep more.” The evidence does not support that logic, and this is the most important finding in this piece. A 2025 umbrella review pooled 28 systematic reviews and meta-analyses covering 118,970 participants and found that digital behavioural insomnia programmes significantly reduced insomnia severity (SMD = -0.42), that the effect was stronger at follow-up (SMD = -0.69), and that it was much larger when a therapist was involved (SMD = -1.05). Within the same body of evidence, an analysis of 52 studies and 12,544 participants found the effect on total sleep time was 0.21 for guided programmes and 0.09 for unguided ones, with the unguided confidence interval crossing zero. In short: these methods reduce insomnia without lengthening sleep. Below is how much evidence stands behind each method, and which long-standing piece of advice no longer holds.
Important note. This article is general information, not medical advice. If your sleep problem occurs more than three nights a week, has lasted longer than three months, or is affecting how you function during the day, see a doctor. Snoring, pauses in breathing, restless legs or excessive daytime sleepiness point to conditions that require separate diagnosis, and nothing here substitutes for that.
What the evidence says is not what most lists say
The most common belief about insomnia is that the problem is simply a shortage: you are not sleeping enough, so you should sleep more. It sounds reasonable, and in practice it turns into advice like “go to bed an hour earlier.”
The largest evidence syntheses describe something different.
Table 1: The effect of digital behavioural insomnia programmes (verified data)
| Measure | Effect size (SMD) | Note |
|---|---|---|
| Insomnia severity (ISI), post-treatment | -0.42 | p < 0.01 |
| Insomnia severity, at follow-up | -0.69 | p < 0.01, effect holds and grows |
| Insomnia severity, with therapist guidance | -1.05 | p < 0.01 |
| Total sleep time | No significant difference between post-intervention and follow-up | – |
Source: Umbrella review and meta-meta-analysis of digital therapeutics for insomnia, NPJ Digital Medicine, 2025. Systematic search of seven databases from inception to October 2024; 28 systematic reviews and meta-analyses, 118,970 participants in total. Methodological quality assessed with AMSTAR 2. Figures as published. ISI is the Insomnia Severity Index; negative values indicate improvement.
Table 2: Is insomnia falling, or is sleep getting longer? (CEOtudent derived analysis)
| Programme type | Effect on insomnia severity | Effect on total sleep time | Ratio of severity effect to sleep-time effect |
|---|---|---|---|
| Guided internet-delivered programme | -0.71 (95% CI -1.18 to -0.24) | 0.21 (95% CI 0.12 to 0.30) | 3.38x |
| Unguided internet-delivered programme | -0.78 (95% CI -1.18 to -0.38) | 0.09 (95% CI -0.03 to 0.20) | 8.67x |
The severity and sleep-time effect sizes are taken as published from an analysis of 52 studies and 12,544 participants included in the umbrella review above. The ratio column was computed by CEOtudent as the quotient of the two absolute effect sizes. The critical detail: for unguided programmes the total-sleep-time confidence interval contains zero, so the increase in that group cannot be statistically distinguished from none at all. Because the ratio compares two different scales, read it as an indication of order of magnitude rather than an exact multiple.
What Table 2 says is direct: these programmes markedly reduce people’s insomnia while increasing their total sleep either very slightly or not at all. The improvement comes from a change in the relationship with sleep.
That conclusion invalidates the most popular item on the older advice lists.
The advice that no longer stands: “go to bed an hour earlier”
Many lists, including the earlier version of this article, recommended getting into bed earlier to fight insomnia. The logic is intuitive: if you cannot sleep, give yourself more chances to.
The evidence points the other way. In the data above, improvement does not come from sleep duration. When someone with insomnia extends the time they spend awake in bed, they keep associating the bed with wakefulness rather than sleep. That is why current behavioural programmes generally work in the opposite direction: they shorten time in bed rather than extend it.
Because that kind of restriction can temporarily increase daytime sleepiness depending on how it is applied, it is a method to run with professional support. We are not giving a protocol here; what we are giving is the reason the old advice pointed the wrong way.
Methods ranked by strength of evidence
A structured behavioural programme (strongest evidence). This is the approach Tables 1 and 2 describe. It is a multi-component programme that addresses how you think about sleep, the association between bed and sleep, and the time spent in bed, together. App-based and internet-delivered versions exist.
Adding guidance (the single largest difference). In Table 1 the effect with therapist guidance is -1.05 against an overall effect of -0.42. That is a difference of 2.50 times. It is the highest-return single decision in this article: run the same content with a professional monitoring it rather than working through it alone.
Durability of the effect. The effect at follow-up (-0.69) is stronger than the effect immediately after treatment (-0.42), a factor of 1.64. That suggests behavioural programmes produce an effect that settles rather than fades once the programme ends. For anyone looking for something lasting rather than a temporary fix, that is significant on its own.
Morning light and darkness at night. That light influences the body clock is well established, and a regular morning light habit is a reasonable step. However, neither of the two sources above measured this practice in isolation, so we are not presenting it as carrying evidence equivalent to the tables. Separately, our chronotype self-test can help you work out which part of the day is naturally yours.
Other “natural” remedies. Herbal teas, supplements and similar suggestions were not evaluated in the two sources this article rests on. Passing judgement on them, favourable or otherwise, on the basis of a source that did not measure them would be wrong, which is why they do not appear here. If you are considering a supplement, discuss it with your doctor because of the possibility of interactions with medication you already take.
The link between sleep and the working day
Insomnia is usually attacked only in the evening hours, yet the structure of the day carries over into the night. A working day that never mentally closed is a working day that continues in bed. How the day ends therefore matters as much as the evening routine; our piece on the shutdown ritual covers that side, and it applies the same discipline of separating where the evidence is strong from where it is thin.
Managing yourself like a CEO has a concrete meaning here: sleep is not a question of willpower, it is a question of a system. The system’s inputs accumulate through the day, and at night you only see the output.
When to see a doctor
If any of the following apply, do not spend time on self-administered methods:
- The sleep problem occurs more than three nights a week and has lasted longer than three months.
- Your daytime functioning, mood or safety is affected, particularly if you drive.
- There is snoring, pauses in breathing at night, waking up gasping, or restless legs.
- The sleep problem began around the same time as a change in medication or an existing health condition.
- It is accompanied by a marked change in mood.
These may require different diagnoses, and a behavioural programme alone may not be enough.
FAQ
Doesn’t sleeping more fix insomnia?
The evidence shows the improvement does not come from sleep duration. In Table 2 the total-sleep-time effect for unguided programmes is 0.09 with a confidence interval crossing zero, while insomnia severity in the same programmes falls markedly. What changes is not the duration, it is the insomnia itself.
Do app-based programmes actually work?
In the review covering 118,970 participants, yes: SMD -0.42 after treatment and -0.69 at follow-up. But the same review shows the effect rising to -1.05 with therapist guidance. An app is a reasonable starting point, but adding guidance produces the largest single measured difference.
Does the effect fade over time?
Not in this data. The effect at follow-up is stronger than immediately after treatment, a pattern consistent with a habit change that settles.
What about melatonin or herbal supplements?
The two sources behind this article did not evaluate them, so we make no judgement here. Because supplements can interact with medication, that decision belongs with your doctor.
Why is this article different from its earlier version?
Because the earlier version carried no sources, was written in the first person, and its most popular item, “go to bed an hour earlier,” pointed in the opposite direction to what current behavioural approaches indicate. When we find that content has expired, we correct it.
Can I apply these methods on my own?
For the general steps, yes, but methods involving a reduction in time spent in bed should be run with professional support because they can temporarily increase daytime sleepiness. That matters especially if you drive or do work requiring sustained attention.
Sources
Umbrella review and meta-meta-analysis of digital therapeutics for insomnia. NPJ Digital Medicine, 2025. Systematic search of seven databases to October 2024; 28 systematic reviews and meta-analyses, 118,970 participants; quality assessed with AMSTAR 2.
Meta-analysis of internet-delivered cognitive behavioural interventions for insomnia, 2023; 52 studies, 12,544 participants. Effect sizes for insomnia severity, sleep quality and total sleep time in guided and unguided programmes. Cited here as reported in the umbrella review above.
AMSTAR 2, a measurement tool for assessing the methodological quality of systematic reviews, used in the umbrella review above.
The ratio column in Table 2 and the comparisons of 2.50 and 1.64 times in the text were computed by CEOtudent from the published effect sizes cited above and recomputed independently before publication.
This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.
This post is also available in:
