You're doing the work and the scale isn't moving. Before anything else: how long has it actually been?
If the answer is under three weeks, there may be nothing wrong at all — skip to the fifth section, because you are probably losing fat and unable to see it yet. If it's been six weeks or more of a genuinely flat trend, one of the following is happening. They're ordered by how often they turn out to be the culprit, not by how comfortable they are.
1. You're eating more than you think
This is the answer most of the time, and it isn't a moral failing — it's a measurement problem that affects everyone, including dietitians.
The research here is consistent and quite brutal: people under-report their intake by roughly 20–30% on average, and the gap tends to widen the more there is to report. Nobody is lying. The calories are in the places nobody counts — the oil the vegetables were cooked in, the handful while cooking, the three bites off a child's plate, the "just a splash" of milk repeated five times a day, the weekend that doesn't get logged because it wasn't a normal day.
A 20% gap on a 2,000-calorie day is 400 calories. That is an entire deficit, erased, invisibly.
What to do: log everything for one honest fortnight, weekends included, before concluding anything else. Not forever — just long enough to see where the gap is. Most people find it in two places and are surprised by both.
2. Your maintenance is lower than the calculator said
Every calorie target starts as an equation fitted to a population, and you are not a population. Genuine individual variation runs to several hundred calories a day, and if the formula guessed high for you, the "deficit" you've been eating may simply be your actual maintenance.
Rest days
1,829kcal
A gentle deficit for steady fat loss. Training adds about 315 kcal on the days you do it — so the two numbers differ, as they should.
This is worth rerunning if you've lost weight since you last did it: a lighter body burns less, so the target that worked at the start of a diet stops working partway through. Lose 10 kg and your maintenance drops by roughly 150–250 calories a day on body size alone.
What to do: measure rather than recalculate. Eat consistently for three weeks, compare your weekly average weight at the start and end, and let the change tell you what your real maintenance is. The maintenance article walks through the arithmetic.
3. You're moving less than you were
A deficit doesn't only reduce what goes in. It quietly reduces what goes out — you take the stairs less, you fidget less, you're slower around the house, your sessions lose a bit of intensity. Unconscious daily movement can fall by a few hundred calories a day, and none of it feels like a decision.
What to do: don't try to out-train it, which usually deepens the fatigue that caused it. Anchor something countable instead — a step target, a walk you take whether or not you feel like it — so the drop has a floor.
4. Exercise calories are being double-counted
Most apps and most watches are generous, and the error compounds in a specific way: the "calories burned" figure for a workout usually includes the calories you'd have burned anyway just being alive for that hour. That resting portion is already inside your maintenance number. Count it twice and a session credited at 500 has really earned you perhaps 350.
Eat that difference back and a five-session week reclaims most of your deficit.
What to do: if you eat back exercise calories, eat back roughly half of what you're shown — or use a system that adds training on top of a baseline which deliberately excludes it.
5. You're reading the scale wrong
Body weight swings by one to two kilos for reasons that have nothing to do with fat: sodium, carbohydrate and the water that travels with it, the state of your digestion, a hard session's inflammation, alcohol, sleep, and — if you have a menstrual cycle — reliably by phase.
Those swings are bigger than a week's fat loss. Someone losing a genuine half-kilo a week will see, on individual mornings, the number go up. Repeatedly. If you're judging progress morning to morning, you will conclude you've stalled several times during a diet that is working perfectly.
What to do: weigh most mornings under the same conditions, then ignore every individual reading and compare weekly averages. Why the scale lies goes into this properly.
Four weeks is the minimum honest window
At a sensible rate you're aiming to lose something like 0.5–1% of body weight a week. For someone at 80 kg that's 400–800 g — well inside the range that daily water shifts can hide. Four weeks of weekly averages is roughly the shortest window in which real progress reliably clears the noise.
6. Sleep, stress and a genuinely long diet
These are real, they are frequently overstated, and they belong last rather than first — because "it must be my hormones" is a comfortable place to stop looking.
Short sleep raises appetite and lowers the quality of the next day's food decisions. Chronic stress does something similar. Neither stops fat loss on its own; both make a deficit much harder to hold, which produces the same result through the front door.
A long uninterrupted diet is its own problem. After several months, adherence is harder, unconscious movement is well down, and the sensible move is often a planned two to four weeks at maintenance rather than cutting further. You don't lose during it. You do come back able to.
If you have symptoms beyond a stalled scale — persistent fatigue, cold intolerance, hair changes, irregular cycles — that's a conversation with a doctor, not an article.
The honest summary
In practice it is almost always the first or second explanation, and almost always found by measuring for a few weeks rather than by changing the plan again. The instinct when the scale stalls is to cut harder, which tends to make intake less accurate, movement lower and adherence worse — all three of the real causes at once.
What Sophia does with this
She works out which one it is, because she holds the pieces separately. She knows what you logged, she knows what your weight actually did, and she compares the two — so when the arithmetic says you're eating several hundred more than you think, she can tell you that rather than guess. She adds training on top of a baseline that excludes it, so it never double-counts. She reads the weekly trend rather than the morning number. And if the trend really has been flat for weeks, she says so, tells you which explanation fits your data, and moves the target instead of telling you to try harder.
If you've been running this loop inside a tracker and it keeps ending at "try harder", that difference — a coach that interprets the stall instead of just recording it — is exactly what metrIQ vs MyFitnessPal walks through.