Two people describe their diet in identical words. One is in ketosis; the other has never been near it. The difference is not willpower and it is not carbohydrate — it is two variables neither of them is watching.
Two people describe their diet in exactly the same words.
Both say they have cut the carbs and upped the fat. Both eat meat, eggs, butter, avocado, olive oil. Both avoid bread, pasta, rice and sugar. If you asked them to write down what they had eaten in a week, the lists would look like siblings.
One is in nutritional ketosis. The other has never been near it, and has spent eight months wondering why the thing that worked so well for everyone online is not working for him.
“Low carb” describes what you removed. It says nothing about what replaced it — and what replaced it determines nearly everything about what happens next.
And a third, more common than anyone admits: low carb, low everything. Take out the rice and do not replace it. Particularly common in women who have spent years being told to eat less. This is under-eating with a respectable name, and it is the version that produces the flattened cortisol, the lost period and the fatigue that gets blamed on the diet being “too extreme” when the actual problem is that it was too small.
Three metabolically distinct situations. One label.
Protein does not behave like fat in this system. A meaningful proportion of the amino acids from dietary protein can be converted into glucose in the liver, through gluconeogenesis. That process is demand-driven rather than simply supply-driven, so it is not as crude as “protein spikes blood sugar” — but it does mean that at high enough intakes, protein can supply enough glucose to keep ketone production suppressed.
Protein also stimulates insulin. Not as sharply as carbohydrate, but meaningfully. And insulin is the brake on ketogenesis.
Which produces the situation that mystifies people: they cut carbohydrate to almost nothing, eat a great deal of chicken and lean beef and protein powder, test their ketones, and find nothing. They conclude they need to cut carbohydrate further. They are cutting the wrong thing.
Protein is the one macronutrient you should be most reluctant to reduce. It preserves muscle while you are losing weight, it carries appetite regulation, and older adults consistently under-consume it.
So “eat less protein to get into ketosis” is technically effective and often a bad trade. If adequate protein is preventing ketosis, adequate protein probably matters more than the ketosis did.
Once fat becomes the majority of your energy intake, the composition of that fat stops being a rounding error. At 25% of calories you can get away with a fair amount of inattention. At 70%, what you chose is most of what you ate.
And this is where the ketogenic conversation goes strangely quiet, because the popular framing — fat is back, saturated fat was never the villain — makes the question feel settled when it is not.
Contested
Saturated fat. Individual response varies more than population data suggests. There is a recognised phenotype, more common in lean and metabolically healthy people, where LDL rises dramatically on a ketogenic diet. Whether that carries the same risk as elevated LDL in other contexts is genuinely unsettled. What is not contested is that you should know whether it is happening to you — which requires a blood test rather than an assumption.
Well supported
Monounsaturated fat. Olive oil, avocado, nuts. The least controversial category in nutrition and the one with the most consistent supporting evidence. If fat is going to be most of your intake, this is the sensible bulk of it.
Omega-3 and omega-6 balance. Almost entirely ignored in ketogenic discussion, and it should not be. A high-fat diet built on seed oils, conventional meat and no oily fish can push the ratio a long way in one direction. Fat-heavy does not mean essential-fat-sufficient.
MCTs behave differently — absorbed and metabolised more directly, producing ketones more readily. Which is why MCT oil raises ketones quickly, and also why it can be used to reach ketosis with less dietary restriction. Either useful, or a way of buying a number, depending on what you were after.
Protein intake differs — one is running gluconeogenesis hard enough to keep ketones suppressed.
Fat composition differs — one is on olive oil, oily fish and nuts, the other on processed meat and cheese.
Total energy differs — one is well-fed, the other quietly under-eating.
Metabolic flexibility differs — a decade of insulin resistance changes how quickly someone adapts to fat oxidation.
Digestive capacity differs — bile flow, enzyme output and gut integrity determine whether a large fat increase is absorbed or simply endured.
Activity differs — repeated high-intensity work has glycogen demands that carbohydrate restriction genuinely impairs.
Know your protein intake. Not roughly — actually. It is the variable most likely to explain why the diet is not doing what you expected, in either direction.
Know what your fat is made of. Roughly what proportion is monounsaturated, what is saturated, and whether any omega-3 is arriving at all.
Get a lipid panel including ApoB, before and after a few months. If you are one of the people whose LDL climbs steeply, you want to know while you can still make a decision about it.
Check the omega-3 index, particularly if fat is the majority of your intake. Cheap, responsive to what you eat, and the marker most likely to be quietly poor on a fat-heavy diet built without thinking about it.
The first article argued that ketosis is an intervention rather than an identity, and that if you are not measuring, you are not doing it.
This one argues something narrower. Even among people genuinely doing it, “low carb, high fat” describes at least three different diets with different consequences — and the two variables that separate them are the two nobody watches.
Protein decides whether it works at all. Fat type decides what it does to you while it does.
Neither is what people argue about, which is roughly how nutrition tends to go.
The answer is in your protein intake and a lipid panel — not in how strictly you avoided the bread.
Ask AIdan → Book a Discovery Call