In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
Every MOMS shake is 16 ounces and every one clears 30 grams of protein. If protein is fixed and volume is fixed, what is left for personalization to actually move - and why does the Light label end up carrying more protein than the Meal Replacement label?
Hypothesis
We expected protein to be the main axis of personalization. That was the working assumption behind the whole system: a Meal Replacement should carry more protein than a Light shake, a Build goal more than a Detox goal, and the Score should climb with it. LAB #006 set a protein floor of roughly 28 grams for meal-replacement formulas, and we assumed the floor would be a floor - a minimum that recipes rose above by different amounts depending on who the shake was for. We expected to find a protein gradient across the four calorie labels. We found a wall instead, and the gradient we did find was pointing the other way.
Method
Moore and colleagues (American Journal of Clinical Nutrition, 2009) fed graded protein doses after resistance exercise and found muscle protein synthesis plateaued at 20 grams of intact protein; larger doses were oxidized rather than used. Mamerow and colleagues (Journal of Nutrition, 2014) ran a 7-day crossover in healthy adults and found that distributing protein evenly at roughly 30 grams per meal stimulated 24-hour muscle protein synthesis more effectively than an isonitrogenous diet skewed toward dinner. Together these describe a per-meal target with a hard bottom and a soft top. Internally, we pulled the macro columns for all 201 recipes in the database, computed the coefficient of variation for each macro across the full set, then grouped every recipe by its Cal Label - Filling (65), Standard (77), Light (45), Meal Replacement (14) - and compared group means, protein ranges, and calorie overlap. We then read individual recipes at the extremes.
Results
Serving size is not a variable: 200 of 201 recipes are 16oz. Neither is protein. Across the whole database protein averages 33.02g with a standard deviation of 2.34g - a coefficient of variation of 7.1%. Not one recipe in 201 falls below 30.0 grams, and 169 of them sit between 30 and 35. The fifth percentile is 30.4g and the median is 32.4g. Every other macro moves freely by comparison: fat has a CV of 38.0% and runs from 2.5g to 22.0g, fiber 32.2% and runs 2.5g to 14.4g, carbohydrate 26.9% and runs 10.5g to 62.0g. Calories are almost as tightly held as protein at 7.2%, capped at 399.9. So the database has two constraints and three free parameters, and protein is a constraint. Which is exactly why the labels invert. The 45 Light recipes average 33.39g protein on 358.66 kcal and 35.96g carbohydrate. The 14 Meal Replacement recipes average 31.83g protein on 390.96 kcal and 43.09g carbohydrate. The Light shakes carry 1.56 grams more protein while carrying 32.30 fewer calories. Twenty-five of the 45 Light recipes reach 33g protein or higher; only 3 of the 14 Meal Replacements do. Meal Replacement has the narrowest protein range of any label, 30.1g to 34.5g - it sits on the floor. The reason is structural. A Light shake has a calorie budget of roughly 360 and still has to clear 30g of protein, so protein takes a large share of a small budget and everything else is trimmed around it: Light recipes have the lowest fiber of any label at 6.94g and the lowest carbohydrate at 35.96g. A Meal Replacement has 391 calories to work with and a mandate to be a whole meal, so it clears the same floor and spends the surplus on carbohydrate and fiber. Two recipes make it concrete. Even Waters (WAY-W1-HYDRATE-008-V1) is labeled Light: 313.2 kcal, 39.2g protein from Chicken 120g, 16.2g carbohydrate, 9.9g fat, 2.5g fiber, Score 16. Golden Charge (EY-W1-ENERGY-008-V1) is labeled Meal Replacement: 397.4 kcal, 30.5g protein from Chickpeas 60g plus Soy Milk 120ml plus Black Beans 35g plus Egg White 150g, 53.2g carbohydrate, 7.8g fat, 11.6g fiber, Score 13. The Meal Replacement has 84 more calories and 8.7 fewer grams of protein than the Light shake. Nothing went wrong in either recipe. Both cleared the floor, and then the calorie label decided what got added on top of it.
What Failed
Two things. First, our own labels do not hold up as macro categories, and the data says so plainly. Filling recipes average 384.23 kcal and 32.80g protein; Meal Replacements average 390.96 kcal and 31.83g protein. That is 6.73 calories and less than a gram of protein apart. The Filling calorie range of 270.5 to 399.9 completely swallows the Meal Replacement range of 366.2 to 399.4, and 54 of the 65 Filling recipes fall inside it. We cannot point at a macro rule that separates Filling from Meal Replacement, because there isn't one - the boundary is an editorial judgment we had been describing as a formulation category. Second, the reward for exceeding the floor is not visible anywhere in our data. Muscle Renew (WAI-W2-PROTEIN-008-V1) carries Chicken 150g for 48.2g of protein, the highest in the database and 60% above the floor, and scores 14 - below the database mean of 14.98. Petal Renewal (WAY-W4-RENEWAL-002-V1) carries 41.5g and scores 18. Across all 201 recipes, protein correlates with Score at r = +0.07, which is nothing. Moore's plateau predicts exactly this: past the point where the dose is used, additional protein is metabolically quiet, and our Score has no reason to notice it. The honest reading is that the top 19 recipes above 36g protein are not better shakes. They are recipes where a whole-food protein source that happened to be dense - almost always chicken at 120-150g - was chosen for texture or flavor and carried protein along with it.
What We Learned
Protein in our system is a constraint, not a dial. It was engineered that way in LAB #006 and then it stopped being interesting, which is the correct outcome for a constraint and a confusing one for a label. Once every shake clears 30 grams, protein cannot also be the thing that distinguishes shakes from each other, and any label implying it does is describing calories and carbohydrate while pointing at protein. What personalization actually moves is fat, fiber, and carbohydrate - the three macros with the room to move - and those are where a body type, a week, and a goal are actually expressed.
What Changed at MOMS
The 30-gram protein floor is now a validated pre-condition rather than a formulation goal: a recipe that fails it is rejected before it is ever scored, and no recipe is credited for exceeding it. Cal Label is being redefined as an explicit calorie-and-carbohydrate rule with non-overlapping bands, because Filling and Meal Replacement currently share both. Any recipe above 38g protein now requires a written reason in the recipe notes - texture, allergen substitution, or amino acid profile - so that dense protein sources are chosen deliberately instead of arriving as a side effect. And customer-facing copy for Light shakes no longer implies less protein, because on our own numbers they average slightly more.
References
1. Moore DR, Robinson MJ, Fry JL, Tang JE, Glover EI, Wilkinson SB, Prior T, Tarnopolsky MA, Phillips SM. Ingested protein dose response of muscle and albumin protein synthesis after resistance exercise in young men. American Journal of Clinical Nutrition. 2009;89(1):161-168. 2. Mamerow MM, Mettler JA, English KL, Casperson SL, Arentson-Lantz E, Sheffield-Moore M, Layman DK, Paddon-Jones D. Dietary protein distribution positively influences 24-h muscle protein synthesis in healthy adults. Journal of Nutrition. 2014;144(6):876-880.
Keep reading
Previous experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type