In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
Every shake in our database lands near 33 grams of protein. So why do the 63 shakes built on soy milk carry 51 fewer grams of eggs, yogurt, beans, tofu and chicken in their protein slot than the 136 that are not?
Hypothesis
In LAB #008 we wrote that the liquid base was "not adding or subtracting protein," because protein totals were matched across every base. Following that logic, we expected the weight of the protein slot to be set entirely by which protein foods a recipe uses. LAB #031 showed that density varies enormously between them. The milk should be irrelevant to it. If soy milk mattered at all, we expected a few grams at most: small enough to disappear inside the differences between egg white, yogurt and chicken.
Method
Published data: Craig and Fresán (Nutrients, 2021) analysed the labels of non-dairy plant-based beverages sold in the United States, Europe and Australia. They found soy- and pea-based drinks averaged roughly 8 to 9 g of protein per serving (240 ml in the US, 250 ml elsewhere). Oat drinks averaged about 2.0 g, nut-based drinks 1.0 to 1.3 g, rice 0.8 g and coconut 0.1 g. Per 100 ml, soy works out to roughly 3.3 to 3.7 g. Internal analysis: we parsed all 201 recipes in the MOMS formulation database. We split them into 63 recipes that use soy milk in the Liquid slot, alone or blended, and 136 that do not. Two recipes list soy milk inside the Protein slot itself, so we excluded them. For each recipe we summed every gram in the Protein slot. To keep ingredient density from driving the result, we then grouped recipes by an identical set of protein ingredients (for example "Egg + Greek Yogurt"). Within each set we compared every soy recipe against every non-soy recipe. That gave 18 matched sets and 198 pairs.
Results
Protein totals were flat: 32.9 g on soy bases vs 33.1 g on non-soy bases. The protein slot was not. It averaged 187.7 g on soy against 238.3 g without, a 51 g gap (median 215 g vs 260 g). Within matched ingredient sets, the soy recipe had the lighter protein slot in 160 of 198 pairs. It tied in 31 and was heavier in 7. The median within-set gap was 30 g. The arithmetic fits. Soy recipes use 184 ml of soy milk on average (range 60 to 334 ml). At the label averages above, that is roughly 6 to 7 g of protein, about a fifth of the shake. Most non-soy bases contribute under 1 g at the same volume. The cleanest pair is WY-W4-RENEWAL-007-V1 "Featherlight Bloom" (Soy Milk 300 ml; Egg 50 g + Greek Yogurt 120 g; 30.2 g protein, 312.9 kcal) against WY-W3-BALANCE-003-V1 "Living Balance" (Almond Milk 279 ml; Egg 50 g + Greek Yogurt 220 g; 31.0 g protein, 337.7 kcal). Same egg, 100 g less yogurt, and only 0.8 g of protein apart.
What Failed
Not all of the 51 g is the milk. The widest pair is EY-W4-CALM-008-V1 "Unwind Fully" (Soy Milk 334 ml; Egg 50 g + Egg White 105 g = 155 g; 30.7 g protein) against EY-W1-ENERGY-011-V1 "Amber Rise" (Almond Milk 220 ml; Egg 90 g + Egg White 195 g = 285 g; 35.1 g protein). Its 130 g gap also carries a 4.4 g higher protein target, so soy explains only part of it. The soy group also uses less egg white, a low-density protein, than the rest of the menu. That mix difference is why the matched-set median (30 g) is well below the raw 51 g. Seven pairs run the other way. WY-W4-RENEWAL-001-V1 "Soft Renewal" (Coconut Milk 150 ml + Soy Milk 150 ml; Tofu 50 g + Greek Yogurt 220 g = 270 g) carries a heavier slot than its non-soy siblings at 260 g. It spent the soy protein on a higher total, 35.2 g, instead of a lighter slot. Nor is this a calorie strategy. The lighter slot saved only 5.6 kcal on average (371.1 vs 376.7 kcal), because soy milk brings its own energy and more fat (11.4 g vs 9.7 g). Our internal Score barely separates the groups (15.3 vs 14.8), so this is not a pattern the Score is detecting. Finally, the per-100 ml figure is a published label average, not a measurement of the specific soy milk we buy.
What We Learned
When the base is soy, the liquid is a protein ingredient. About 184 ml of soy milk does the work of roughly 30 to 50 g of solid protein food. LAB #008 was wrong to say the base was "not adding or subtracting protein." Totals matched only because the protein slot had already compensated. The trade soy milk makes is weight and allergen coverage, not calories. Every soy-base recipe in the database carries a soy declaration.
What Changed at MOMS
The formulation sheet now logs protein from the liquid base as its own line, using the label value of the specific milk we stock, before the protein slot is sized. When an allergy screen removes soy from a recipe, the protein slot is re-sized upward, typically by 30 to 50 g of solid protein food, rather than treating the base swap as neutral. We are also treating the "not adding or subtracting protein" line in LAB #008 as corrected by this note.
References
Craig WJ, Fresán U. International Analysis of the Nutritional Content and a Review of Health Benefits of Non-Dairy Plant-Based Beverages. Nutrients. 2021;13(3):842.
Keep reading
Not sure which shake fits you? We match the formulation to your body type.
Find your body type