Enter the same height, weight, age, sex and activity level into four fitness apps and you will get four different protein targets, sometimes sixty grams apart. At most one of them can be right, and probably none of them is right for you specifically.
General nutrition information, not medical advice. If you are managing a health condition, pregnant, or under 18, speak to a doctor or dietitian.
Where the number comes from
Essentially every app does the same three steps, and the disagreements come from small choices inside each one.
- Estimate your resting metabolic rate from a formula — Mifflin–St Jeor, Harris–Benedict, or Katch–McArdle if it asked for body fat. These disagree with each other by a couple of hundred calories.
- Multiply by an activity factor, typically 1.2 to 1.9. This is the big one: the gap between “lightly active” and “very active” can be six hundred calories, and you are the one guessing which you are.
- Split the result into macros, either by percentage of calories or by grams per kilogram. Percentage splits produce the strangest results, because the protein target then moves when you change your calorie goal — which is backwards.
Three places it goes wrong
The formulas have real error bars. Mifflin–St Jeor is about as good as these get and is still roughly ±10% for an individual. On 2,400 calories that is a 480-calorie spread. The formula is describing a population; you are one person, with your own thyroid and history.
Activity multipliers are self-reported and inflated. People consistently overestimate. Four gym sessions a week is usually “moderately active”, not “very active”, because an hour of training is a small part of a day spent mostly sitting. One step up the wrong rung and the whole plan is built on a number several hundred calories too high.
Percentage-based macro splits break on edge cases. “30% of calories from protein” gives a very lean, very active person a sensible target and gives a heavy person on a steep deficit a nonsensical one. Protein need scales with bodyweight and training, not with how many calories you happen to be eating this month.
Why it matters less than it seems
Here is the part that gets lost. Being 15% off your theoretically ideal protein target is a much smaller problem than missing whatever target you set, three days a week.
- 160 g a day, hit consistently, averages 160 g.
- 185 g a day, hit four days in seven, averages about 106 g.
The second target is closer to optimal and the first outcome is better by a wide margin. Precision in the target is worth very little; consistency against it is worth nearly everything. Which means the useful question is not “is this number exactly right” but “can I actually eat this, every day, for months”.
A target you meet is better than a target that is correct. Apps compete on the second and users fail on the first.
What to do instead
- Take the app’s number as a starting hypothesis, not a fact.
- Pick the activity level one step below the one you want to pick. It is more often right.
- Set protein from bodyweight, not from a percentage. Around 1.6 g/kg if you train, higher if you are dieting. It should not move when your calories do.
- Give it three or four weeks and watch what happens to your weight and your training, then adjust. That feedback beats any formula, because it is measuring you rather than a population you resemble.
- Then make it easy to hit. This is the step everyone skips, and it is the only one that determines whether any of the rest happened.
Protn lets you enter your own targets for exactly this reason, and it will estimate them if you would rather it did — but the work it actually does is step five: turning a number into a menu you can follow without deciding anything. Have a look.