In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
Chia and flaxseed sit in our formulas as fiber ingredients. Across all 201 recipes, does the functional slot actually predict how much fiber a shake ends up with - or is something else doing that work?
Hypothesis
We expected a clear positive relationship. Chia seed is 34 to 40 grams of dietary fiber per 100 grams (Khalid et al., 2023), which is denser in fiber than any produce item we use. At our median functional-slot weight of 7 grams, chia alone should carry 2.4 to 2.8 grams of fiber - roughly a third of a typical MOMS shake's total. Flaxseed is not far behind. If the slot does what its name implies, functional-slot grams should track total fiber across the database. We also expected fat to move, but assumed it would be the smaller effect of the two.
Method
Published research. Khalid et al. (2023), a review in Food Science and Nutrition, reports chia seed at 34 to 40 g dietary fiber per 100 g. Novotny, Gebauer and Baer (2012), a controlled 18-day feeding trial in 18 adults published in the American Journal of Clinical Nutrition, measured the metabolizable energy of almonds at 4.6 kcal/g against the 6.0 to 6.1 kcal/g predicted by Atwater factors - a 32 percent overestimate. We used the first to set a fiber expectation for the slot and the second as a standing caution that the gram weight of a nut or seed is a poor guide to what it actually delivers. Internal analysis. We parsed all 201 recipes in the MOMS database into their six ingredient slots (Liquid, Body, Fruit and Vegetable, Protein, Functional, Flavor) and converted every entry to grams. Recipes were grouped three ways by functional-slot contents: no functional slot at all (n=77), a nut or seed present (n=81), and spice or citrus only (n=43). We then computed Pearson correlations of each recipe's total fiber against functional-slot grams, against grams of legume in the protein slot (chickpeas, black beans, lentils, adzuki, tofu), and against grams in the produce slot. Three recipes were pulled for a line-by-line read: EI-W2-WARM-005-V1, EI-W1-CLEANSE-005-V1 and EY-W3-GUT-004-V1.
Results
The functional slot is the smallest thing we put in a shake. Median weight 7 g, range 0.2 g to 26.8 g, against a median total shake weight of 592 g. It is 1.2 percent of the glass. That 1.2 percent moves fat hard. The 81 recipes carrying a nut or seed in the functional slot average 12.6 g fat. The 77 recipes with no functional slot at all average 7.8 g. A difference of 4.8 g, or 62 percent, out of a median seven grams of food. Correlation of fat against functional-slot grams: r = 0.55. It moves fiber not at all. Those same 81 nut-or-seed recipes average 7.8 g fiber. The 77 with no functional slot average 7.8 g. Correlation of total fiber against functional-slot grams across all 201 recipes: r = 0.001. The slot that does predict fiber is the protein slot. Correlation of total fiber against grams of legume in the protein slot: r = 0.52. Sorted into thirds by legume weight, the bottom third averages 7 g of legume and 6.2 g of fiber, the middle third 45 g and 7.4 g, the top third 73 g and 9.2 g. The 51 recipes with no legume protein at all average 5.8 g fiber; the 150 with legumes average 8.2 g. Produce grams, for comparison, correlate with fiber at r = 0.07. Read at the recipe level it is starker. EI-W2-WARM-005-V1 (Ember Boost, Filling) carries Turmeric 1.2 g + Chia Seeds 20 g. Twenty grams of chia carries 6.8 to 8.0 g of fiber on its own. Ember Boost finishes at 10.5 g fiber, 364.6 kcal. EI-W1-CLEANSE-005-V1 (Calm Clean) is the same body type, the same Filling label, five calories away, and has no functional slot whatsoever - and finishes at 9.3 g fiber. Twenty grams of chia bought 1.2 g of net fiber. Calm Clean gets to 9.3 g on Black Beans 45 g + Chickpeas 30 g in the protein slot. Where the chia did show up was fat: 15.2 g in Ember Boost against 8.2 g in Calm Clean.
What Failed
The headline number is true of the slot and false of chia, and we should say so plainly. Split the 81 nut-or-seed recipes apart and they do not behave alike. The 25 recipes containing chia average 9.2 g fiber - 1.4 g above the no-slot baseline of 7.8 g. The 19 containing flaxseed average 8.2 g. What flattens the category to zero is the tree nuts: the 40 recipes whose functional slot is cashew, pecan or walnut with no seed present average 6.9 g fiber, which is 0.9 g BELOW the shakes with no functional slot at all. Cashew is roughly 3 percent fiber and 44 percent fat. Our r = 0.001 is an honest number for a bucket that should never have been one bucket. The second failure is the design itself. This is 201 different recipes compared against each other, not one recipe with the seed slot swapped in and out. The functional slot co-varies with everything around it - the 81 nut-or-seed recipes also run 6.2 g lower in carbohydrate, which means their produce and liquid choices differ too. We cannot say the seeds caused the fat gap, and we cannot say the legumes caused the fiber. All we can say is that in the recipes we actually ship, functional-slot grams track fat and do not track fiber, and legume grams track fiber. A controlled single-recipe swap would be the real test and we have not run one.
What We Learned
Name a slot for what it delivers, not for why you reached for it. We called this the functional slot because chia and flax are what you add when you want fiber. In practice it is our fat-and-aroma slot: a median seven grams that moves fat by 62 percent and, as a category, moves fiber by nothing. Fiber is decided two slots upstream, in the protein choice, and it is decided before the functional slot is ever filled. A shake built on egg white and chicken starts near 5.8 g of fiber. No realistic amount of chia closes that gap: seven grams of chia is about 2.4 g of fiber, while 45 g of black beans is more than that and also happens to be the protein you already needed. The cheapest fiber in a whole-food shake is the fiber that arrives attached to something else you were adding anyway.
What Changed at MOMS
We stopped treating chia and flax as the fiber lever. Fiber is now a protein-slot decision: any formulation that models below 7 g of fiber gets reworked at the protein slot first, swapping part of the egg white or chicken for chickpeas, lentils or black beans, before anyone touches the functional slot. The functional slot is now specified against a fat target and a flavor or aroma role, which is what it was actually doing all along. And we split it in our own internal sheets: seed entries (chia, flaxseed) and nut entries (cashew, pecan, walnut) no longer sit under one heading, because in our data they move fiber in opposite directions and averaging them together hid that for months.
References
Khalid W, Arshad MS, Aziz A, Rahim MA, Qaisrani TB, Afzal F, Ali A, Ranjha MMAN, Khalid MZ, Anjum FM. Chia seeds (Salvia hispanica L.): A therapeutic weapon in metabolic disorders. Food Science & Nutrition, 11(1), 3-16, 2023. | Novotny JA, Gebauer SK, Baer DJ. Discrepancy between the Atwater factor predicted and empirically measured energy values of almonds in human diets. The American Journal of Clinical Nutrition, 96(2), 296-301, 2012.
Keep reading
Previous experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type