In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
When the body-type code changes, does the shake's calorie target move - or only the macronutrient carrying those calories?
Hypothesis
We expected body type to act as a second energy dial. The heavier, warming codes would be built larger and the lighter codes smaller, with total calories doing most of the personalization work on top of whatever the Cal Label had already set. If that were true, grouping the database by body type should show a clear calorie gradient across the eight codes, with the macros following along behind it.
Method
We started from the published evidence. Gardner and colleagues (JAMA, 2018) randomized 609 overweight adults to a healthy low-fat or a healthy low-carbohydrate diet for twelve months and found no significant difference in weight change - 5.3 kg lost on low-fat versus 6.0 kg on low-carbohydrate - and no interaction with genotype pattern or insulin secretion. If macronutrient allocation at matched diet quality does not reliably change the energy outcome, then allocating macros by body type cannot be justified as an energy decision, and it should not be built as one. Internally we analyzed all 201 recipes in the production database. Each recipe's BT code was collapsed to its body-type family (E, W, M, WA), and every comparison was stratified by Cal Label - Filling, Standard, Light, Meal Replacement - so that the existing energy label could not confound the result. We computed family means for Kcal, Protein, Carb, Fat and Fiber within each stratum, the Pearson correlation between carbohydrate and fat inside each stratum, and then pulled matched recipe pairs from opposite ends of the fat axis that shared both a Cal Label and a Score.
Results
Calories barely move. Inside the Standard band (n = 77) the four families land at 372.5 kcal (E, n = 25), 373.8 kcal (W, n = 20), 371.9 kcal (M, n = 9) and 376.0 kcal (WA, n = 23) - a spread of 4.1 kcal, about 1.1 percent. Across those same four groups carbohydrate runs from 34.0 g (WA) to 46.0 g (M), a 12.0 g or 35 percent spread, and fat runs from 6.5 g (M) to 12.1 g (WA), a 5.6 g or 86 percent spread. Protein sits nearly still at 32.1 to 34.6 g. The substitution is reciprocal rather than additive: carbohydrate and fat correlate at r = -0.71 across all 201 recipes, -0.73 within Standard, and -0.95 within Meal Replacement (n = 14). Two matched Filling recipes make it concrete. WAY-W1-HYDRATE-011-V1 "Quenched Core" - Almond Milk 300 ml, Mango 50 g + Avocado 35 g, Chicken 80 g + Egg 28 g, Cacao 5 g + Pecan 9.8 g, Agave 5 g - delivers 391.3 kcal at 19.8 g carbohydrate and 22.0 g fat. MY-W2-DETOX-011-V1 "Wellspring Reset" - Rice Milk 143 ml, Cauliflower 45 g, Spinach 30 g + Dates 21.1 g, Egg 50 g + Chickpeas 60 g + Egg White 150 g - delivers 391.8 kcal at 49.9 g carbohydrate and 8.3 g fat. Same energy label, same Score of 16, 0.5 kcal apart, and 30.1 g of carbohydrate and 13.7 g of fat apart. The arithmetic closes: 30.1 g of carbohydrate at 4 kcal per gram is plus 120 kcal, 13.7 g of fat at 9 kcal per gram is minus 123 kcal, and 2.0 g of protein is minus 8 kcal, predicting a 11 kcal deficit against an observed surplus of 0.5 kcal - an 11 kcal residual, 2.8 percent of the meal, from ingredient-level rounding. A Standard pair repeats the pattern at wider amplitude: WAI-W4-RENEWAL-001-V1 "New Leaf" (389.6 kcal, 11.0 g carbohydrate, 21.1 g fat) and MI-W2-IMMUNE-001-V1 "Vital Flow" (389.7 kcal, 50.5 g carbohydrate, 7.6 g fat) sit 0.1 kcal apart across a 39.5 g carbohydrate gap.
What Failed
The calorie constancy is partly an artifact of a ceiling, not evidence of active balancing. Every recipe in the database falls between 270.5 and 399.9 kcal, and 75 of 201 (37 percent) sit at 390 kcal or above, pressed against what is effectively a 400 kcal cap. When the formulation space is bounded that tightly, family means will look stable whether or not anyone is holding them there. We can say the reallocation is isocaloric. We cannot say from this data that isocaloric was anyone's intent. The larger failure is that fiber is not part of the seesaw, and it loses. Fiber tracks carbohydrate (r = +0.52) and moves against fat (r = -0.22), so the fat-forward families arrive lower on fiber inside the same energy band: 6.8 g for WA and 6.2 g for M in the Standard band, against 8.3 g for E and 8.4 g for W. Across the whole database 32 of 201 recipes sit under 5 g of fiber, and 24 of those 32 come from the M and WA families, which together are only 44 percent of the database. The worst case is our own headline recipe: "New Leaf" carries 2.9 g of fiber against "Vital Flow" at 7.5 g, for the same calories. Reynolds and colleagues (Lancet, 2019) put the target at a minimum of 25 to 29 g of fibre per day, with observational data showing a 15 to 30 percent lower rate of all-cause and cardiovascular mortality in the highest versus the lowest fibre consumers. A 2.9 g breakfast is not a neutral trade. Our Score did not catch it - "New Leaf" scored 18, the higher of the pair. The Score is not tracking fiber adequacy, and we should not imply that it is.
What We Learned
Body type is a composition instruction, not an energy instruction. The energy decision has already been made one column earlier, by Cal Label. The BT code only decides which macronutrient carries it. That is a cleaner and more defensible mental model than "some people need bigger shakes," and it matches the trial evidence: if allocation does not reliably change the energy outcome, then allocation has to be judged on the things that ride along with it - satiety, tolerance, preference, and what comes attached to the macro you chose. What comes attached is fiber, and we were not watching it.
What Changed at MOMS
We split the constraint in two. The calorie target now comes only from Cal Label, and the BT code is documented internally as a carbohydrate-to-fat allocation rule rather than a portion rule. That stops formulators from quietly adding energy whenever they move a recipe toward the fat-forward codes, which is the failure mode the old wording invited. We also added a fiber floor that applies independently of body type. No 16 oz shake ships below 5 g of fiber, and any WA or M recipe that lands under 6 g has to recover it in the produce or functional slot - chia, ground flax, or a whole-fruit swap for part of the fat - before it can be scored. The 32 recipes currently under that floor, "New Leaf" among them, are queued for reformulation rather than grandfathered in. The Score itself gets a fiber term in the next revision, because a formulation rule the Score cannot see is a rule that will drift.
References
Gardner CD, Trepanowski JF, Del Gobbo LC, et al. Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion: The DIETFITS Randomized Clinical Trial. JAMA. 2018;319(7):667-679. | Reynolds A, Mann J, Cummings J, Winter N, Mete E, Te Morenga L. Carbohydrate quality and human health: a series of systematic reviews and meta-analyses. Lancet. 2019;393(10170):434-445.
Keep reading
Previous experiment
Next experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type