You already know why you do it. You check your phone because you're anxious, not because you need to. You snap at people you love when you're tired, not because of anything they did. You've had the insight — sometimes for years — and the behavior is still there. This is one of the most consistent findings about how people actually change: understanding a behavior and changing a behavior are two different skills, run by two different systems.
The loop that's actually running the show
Habitual behavior, good or bad, tends to run on a simple loop: a cue triggers a routine, which delivers some kind of reward, which reinforces the cue-routine connection for next time. Once that loop is established, it largely runs below the level where insight lives.1 Insight sits in the part of the mind that explains, narrates, and reflects. The loop sits somewhere that responds to triggers directly, mostly without waiting for permission from the reflective part of you.
This is why understanding a habit's origin rarely dissolves it. Knowing your late-night snacking is about loneliness rather than hunger doesn't interrupt the loop, because the loop was never waiting on that information — it was responding to a cue (a quiet apartment at 11pm) with a routine (opening the fridge) for a reward (a few minutes of distraction). The insight is true. It's just answering a different question than the one the habit is asking.
Why willpower isn't the fix either
The other common approach is trying harder in the moment: gritting your teeth against the cue. It works sometimes, and it fails predictably. Resisting an urge takes attention and effort, and both are in short supply when you're stressed, tired or upset, which is usually when the cue turns up in the first place. Relying on willpower means betting on your best self being available at your weakest moments. That's a bad bet, not a character flaw.
A caution on the science: the popular idea that willpower works like a muscle that gets used up has not held up well. A large multi-lab replication found an effect close to zero.2 What does hold up is simpler. People whose lives are arranged so that the good behavior is easy report fewer lapses than people who rely on resisting temptation.
Where the actual lever is
If the loop runs on cue → routine → reward, there are only two places to intervene that don't require winning a willpower fight every single time: change what's available at the routine step, or change how easy the cue is to reach in the first place.
Take doomscrolling. The cue might be picking up the phone during any idle moment. Fighting the urge to open the app in that moment is the hardest possible place to intervene — it's peak temptation, minimum resources. Redesigning the environment is easier: log out of the app so opening it requires a deliberate re-login, move the phone to another room during the hours it's usually reached for, or put something else — a book, a notebook — physically where the phone used to sit, so the routine step defaults to something else without a fight.
You don't beat a habit loop with more willpower. You redesign it so willpower isn't the thing being asked to do the work.
Good habits use the same mechanism
This cuts both ways — it's also why "just decide to exercise more" so rarely sticks, and why putting your running shoes by the door the night before does. You're not becoming a more disciplined person; you're making the cue impossible to miss and the routine the path of least resistance. The habit that survives isn't the one you wanted most — it's the one that was easiest to start.
What insight is actually for
None of this makes insight useless. Understanding why a habit exists tells you where to intervene — which cue, which reward, which unmet need is being served badly. It just can't do the intervening itself. That part happens in the environment and the routine, not in the explanation.