Five weeks with no scale movement at month five: how do I tell adaptive thermogenesis from adherence drift?

Month five, semaglutide 1.7 mg, and the scale has not moved in five weeks. Before that I was averaging about 0.5 kg a week fairly reliably. I am now at 88.1 kg from a start of 103.4 kg.

Every explanation I read falls into one of three buckets and they imply completely different responses:

1. Adaptive thermogenesis. My metabolism has downregulated and I need to accept a lower expenditure or do something to counter it.
2. Measurement error. I am losing fat but water and glycogen are masking it, and the plateau is an artefact of what a bathroom scale measures.
3. Adherence drift. I am eating more than I was and more than I think, and the honest answer is that my logging has degraded.

I cannot distinguish between these from the inside, and each one has a different fix: eat less, wait, or log properly. What I would like is a way to actually diagnose which of them applies rather than picking the flattering one.

Data I have: daily morning weights, a logged intake averaging 1,480 kcal, a waist measurement I have taken twice, and nothing else. I lift twice a week, walk a lot. No tracked steps.

Emphasising the water-masking case, because it is the one that leads people to make bad decisions and it is the easiest to rule in or out.

Things that add or hold 1-3 kg of water and can hide a month of fat loss:

• New or increased resistance training. Muscle damage draws fluid into the tissue. Starting or intensifying a programme reliably produces 1-2 kg of retention over the first two to four weeks. OP lifts twice a week; if that started or changed around when the plateau began, that is your answer.
• Sodium and carbohydrate. A shift of 100 g/day in carbohydrate intake moves 1-1.5 kg of glycogen-bound water within days.
• Luteal phase. 1-2 kg cyclically, which for a person losing 0.5 kg/week can flatten a two-week window entirely.
• Cortisol. Poor sleep, work stress, high training volume, or all three, via aldosterone. This is the sneaky one because it is chronic rather than episodic.
• Constipation. Very common on these drugs, and 1 kg of retained stool weighs 1 kg. People underestimate this considerably.
• Creatine, if recently started. 1-2 kg of intracellular water in the first fortnight.

The tell is a divergence between weight and every other measure. If your waist is down 2 cm, your clothes fit differently, your ring is loose, and progress photos look different, but the scale is flat, you are not in a stall, you are in a measurement problem, and the correct action is to do nothing at all.

The failure mode this prevents is the common and destructive one: a person in an unrecognised fluid plateau cuts intake by 300 kcal, then the fluid resolves and the scale drops 2 kg in four days, and they conclude that cutting worked. They now believe they need 1180 kcal to lose weight, which is false, and they have locked in an unnecessary restriction that will cause problems in month eight.

Practical rule: never change anything on the basis of scale weight alone, and never on less than three weeks of rolling-average data.

Example of why daily weights mislead:
  daily weights, week 1: 88.4 88.0 88.6 87.9 88.3 88.8 88.1  -> mean 88.30
  daily weights, week 4: 88.2 87.6 88.4 87.5 87.9 88.5 87.6  -> mean 87.96
  headline reading: "stuck at 88"
  actual reading:   -0.34 kg over 3 weeks, i.e. slow but real

Adding a note on the adaptive-thermogenesis argument, because it gets used as an excuse in both directions and the honest position is narrower than either camp claims.

What the evidence supports: after substantial weight loss, measured total energy expenditure sits somewhat below what body-composition equations predict, the effect is detectable years later in some cohorts, and it is partly mediated by reduced leptin signalling with downstream effects on thyroid hormone, sympathetic tone and skeletal muscle efficiency.

What the evidence does not support:

• That the effect is large enough to prevent further loss. A 100-250 kcal/day residual is real and is also about one tablespoon of oil plus a handful of nuts. It changes how much deficit you need to create, not whether creating one works.
• That it is permanent and irreversible. It attenuates with weight stability and it is not a ratchet.
• That "starvation mode" prevents weight loss at low intakes. The thermodynamics do not permit that. What low intakes do produce is spontaneous reductions in unconscious activity, which lowers expenditure by a genuinely large amount and is often mistaken for metabolic magic. That reduction is behavioural and measurable, and it is why maintaining step count during a deficit matters more than most nutritional fine-tuning.

The practical upshot: adaptive thermogenesis is a reason to expect the deficit to need recalculating every 10-15 kg, and a reason to prefer adding expenditure over subtracting intake once intake is already low. It is not a reason to stop, and it is not evidence that anything has broken.

The simplest explanation is usually right and in this case it is that you are eating more than you are logging. Everyone in these threads wants it to be metabolism and it almost never is. Weigh your food for a fortnight and the plateau will explain itself.

The reason I would still not jump straight there: OP is five months in on a drug whose effect on intake is not constant over time, so "eating more" and "the drug is doing less" are the same observation with different implications for what to do about it. But the measurement comes first either way.

To be more useful than a one-line diagnosis, here is why the drift explanation deserves its status as the default rather than as an accusation. Every mechanism that produces it is invisible to the person it happens to:

• Portion creep. The plate you serve yourself gets 10% larger over eight weeks and you cannot perceive a 10% change in a portion. Nobody can.
• Cooking oil. A tablespoon is roughly 120 kcal and it is poured, not measured. Two tablespoons a day, unlogged, is 240 kcal.
• Weekday-weekend asymmetry. Most people log Monday to Thursday carefully and the weekend loosely. Two loose days can add 400 kcal to a weekly average without any single day looking wrong.
• Foods that stopped being logged. Milk in coffee, a child's leftovers, tastes while cooking, a handful of something while making dinner. Each is negligible and the sum is not.
• The item you log by brand rather than by weight, where the actual serving has drifted from the nominal one.

None of that is dishonesty and framing it as such is why people resist the explanation. Framed as measurement error in a self-reported instrument, which is exactly what it is, it becomes something you fix with a kitchen scale for a fortnight rather than something you feel bad about.