Most questions people bring to a search box at midnight have no definition behind them. This one does, and knowing it changes what happens next, because the thing that gets treated is not "bad sleep" β it is a specific pattern with a frequency, a duration and a visible cost. Three nights a week. Three months. Something it is demonstrably taking from you: the mornings, the temper, the work, the patience with people you like. Fail one of those and you probably have a sleep problem rather than insomnia β and sleep problems and insomnia do not come apart the same way.
What is usually going on
It started recently and something started it. A bereavement, a deadline, a newborn, a move, an illness. Short-run sleeplessness with an obvious cause is extremely common and usually ends when the cause does. The useful thing is to interfere with it as little as possible, because the measures people reach for in week two β earlier bedtimes, longer lie-ins, a drink to get over the edge β are most of what turns a short problem into a long one.
It outlived whatever started it. This is the version with a name and a treatment. The original cause is gone and the sleeplessness stayed, because the bed learned something in the meantime: hours of lying there alert, repeated, teach the body that this is a place where you are awake. By then the problem maintains itself and effort feeds it. The standard treatment is mostly the undoing of that learning, and it outperforms what most people expect of something with no prescription attached.
Your clock runs late and you are fighting it rather than failing at it. If you fall asleep at two and wake at ten with no difficulty whenever nothing forces the hour, your sleep works. It is simply not on the schedule you are being held to. Calling that insomnia is both useless and quietly demoralising, and the difference is worth reading properly.
You are not giving it the hours. Six hours of opportunity cannot yield seven hours of sleep. This one is arithmetic rather than pathology, it is far more common than it should be, and it is the only version that improves the same week you admit to it.
What to do tonight
- Start a plain record and keep it two weeks: the hour you got in, your best guess at the hour you went under, the wakings, the hour you got up. Two weeks of rough notes beats any amount of recall, and it is what a clinician asks for first anyway.
- Choose the hour you get out of bed and hold it, including after a wrecked night. Of everything available, that is the lever with the widest reach, and it works precisely because it is dull.
- Stop recruiting effort. Sleep is the one task where trying lowers output. Twenty minutes awake and you leave the bed β not as a penalty, as a way of protecting what the bed means.
- Do not start rationing your hours in bed to "earn" deeper sleep. There is a real technique shaped like that, it belongs with someone supervising it, and improvised alone it usually backfires.
What it depends on
Insomnia gets described as one condition and behaves like three. The kind that fails at the beginning runs on a mind that has not come down yet, and responds to what the hour before bed looks like; the time it takes you to go under is the number that reads it. The kind that fails in the middle runs on fragmentation and asks a different set of questions. The kind that ends your night at half four behaves differently again and is the one most worth raising with a doctor rather than a page. The insomnia check tells the three apart and weighs the daytime cost β the half of the definition people skip, and the half that decides whether any of this needs treating at all.