In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
Two of our shakes list carbohydrate 0.2 grams apart. How much of each number is fiber, and does subtracting it change which shake is actually lower in digestible carbohydrate?
Hypothesis
We expected fiber to be a small and roughly constant share of the carbohydrate number - call it ten percent - which would make total carbohydrate a safe proxy for digestible carbohydrate across the whole menu. Under that assumption the reframe is pedantry: two shakes listing 25 grams of carbohydrate are two shakes with 25 grams of carbohydrate, and the fiber line underneath is a bonus rather than a correction. We also expected, going in, that a higher fiber share would mean a gentler glucose response, and that we would be able to say so.
Method
Published research first. Jenkins and colleagues (American Journal of Clinical Nutrition, 1981) fed 62 commonly eaten foods and sugars to groups of five to ten fasting volunteers, standardizing every challenge on 50 grams of available carbohydrate - total carbohydrate minus fiber - calculated from food tables. Available carbohydrate, not total carbohydrate, is the unit the whole method is built on. Matched that way, the 62 foods produced glycemic indices ranging from 15 percent for soya beans to 97 percent for parsnips. Meng, Matthan, Ausman and Lichtenstein (American Journal of Clinical Nutrition, 2017) held a white-bread challenge fixed at 50 grams of available carbohydrate in 20 to 22 adults aged 50 to 80 and added defined amounts of each macronutrient on top, including 4.8 and 9.6 grams of fiber from oat cereal, sampling arterialized venous blood for two hours. Internal analysis. We took all 201 recipes in the production formulation sheet and computed, for each, the fiber share of the carbohydrate number and net carbohydrate as Carb minus Fiber. We then bucketed recipes by rounded total carbohydrate to find pairs that list the same carbohydrate figure, correlated our internal Score against both fiber grams and fiber share, and re-ranked the entire menu by net carbohydrate to see how far individual recipes moved.
Results
Fiber is a median 19.0 percent of the carbohydrate number across 201 recipes, mean 20.1 percent, with a range from 7.7 percent to 55.2 percent. Fiber itself spans 2.5 to 14.4 grams. Total carbohydrate spans 10.5 to 62.0 grams; net carbohydrate spans 4.7 to 51.2 grams. The matched pair is close to laboratory-clean. EI-W2-WARM-005-V1 "Ember Boost" lists 25.4 g carbohydrate, 10.5 g fiber, 15.2 g fat, 31.6 g protein, 364.6 kcal. WAY-W4-RENEWAL-008-V1 "Quiet Rebuild" lists 25.2 g carbohydrate, 3.0 g fiber, 15.3 g fat, 37.2 g protein, 392.7 kcal. Both are 16 oz, both are labeled Filling, both carry Egg at roughly 27.5 g and Chickpeas at 30 g, and their fat figures sit 0.1 g apart. Their carbohydrate figures sit 0.2 g apart. Their digestible carbohydrate sits 7.3 g apart: 14.9 g against 22.2 g. Quiet Rebuild delivers 49 percent more absorbable carbohydrate than the shake it matches on the label. The ingredient lists show where it goes. Ember Boost's carbohydrate arrives attached to fiber: Chia Seeds 20 g, Pumpkin 40 g, Carrot 30 g. Quiet Rebuild's arrives as White Rice 19.5 g and Agave 9.8 g, both close to fiber-free. A second pair repeats the pattern higher up the range. EY-W1-ENERGY-005-V1 "Bright Start" lists 55.6 g carbohydrate with 13.4 g fiber, for 42.2 g net. MI-W3-VITAL-008-V1 "Bright Vitality" lists 56.3 g with 5.8 g fiber, for 50.5 g net. Both carry an internal Score of 14. Their labels sit 0.7 g apart and their digestible carbohydrate sits 8.3 g apart. At the extremes the composite gets stranger. EY-W4-CALM-008-V1 "Unwind Fully" lists 10.5 g carbohydrate of which 5.8 g is fiber - 55 percent - leaving 4.7 g net. MI-W2-IMMUNE-008-V1 "Resilient Current" lists 45.7 g of which 3.5 g is fiber, leaving 42.2 g. Both are 16 oz and both are labeled Light. Their digestible carbohydrate differs by a factor of nine.
What Failed
Three things, and the biggest is the claim we wanted to make. Our Score does not track this at all. Across 201 recipes the correlation between Score and the fiber share of carbohydrate is -0.007, which is indistinguishable from zero. The correlation between Score and fiber grams is -0.17, faintly negative, meaning higher-Score recipes carry marginally less fiber rather than more. Whatever the Score is weighting, this is not in it, and we are not going to imply otherwise. The blood-sugar claim does not survive the literature. We wanted to write that more fiber inside the carbohydrate number means a gentler glucose response. Jenkins 1981 found no significant relationship between the fiber content of its 62 foods and their glycemic index (r = -0.220, not significant) - the paper that founded the field reported fiber grams failing to predict the response. Meng 2017 added 4.8 and 9.6 grams of fiber to a fixed available-carbohydrate load and found no significant effect on glucose area under the curve, measured meal glycemic index, meal glycemic load, or insulin. So subtracting fiber gives a more honest count of what is absorbed. It does not license a statement about how fast, and we are not making one. And the reframe barely matters where people actually look. Of the 30 recipes lowest in total carbohydrate, 27 are still among the 30 lowest by net carbohydrate - only three swap out. Anyone who simply wants our lowest-carbohydrate shake gets there fine using the listed number. The reordering happens in the middle of the menu, where 27 of 201 recipes move more than 20 places.
What We Learned
The carbohydrate number on a whole-food shake is a composite, not a quantity. A median fifth of it is fiber: material counted as carbohydrate by labeling convention and not absorbed as carbohydrate by the person drinking it. That makes total carbohydrate a weak unit for comparing two of our shakes against each other, because the fiber share swings from 8 percent to 55 percent depending on whether the carbohydrate came in as chia and pumpkin or as white rice and agave. It remains a perfectly good unit for comparing a whole-food shake against a powder-and-water one, where the fiber share sits near zero. But the correction is an accounting correction. Knowing that 25 grams is really 15 tells you what crosses the gut wall. It does not tell you how quickly.
What Changed at MOMS
Net carbohydrate - Carb minus Fiber - now sits as a derived column beside total carbohydrate in the formulation sheet, and formulation review compares recipes on the net figure rather than the listed one. We added one check to the rotation build. When two recipes serve the same Goal and land within 1 gram of each other on total carbohydrate, they are flagged if their net carbohydrate differs by more than 5 grams. Ember Boost and Quiet Rebuild would trip that flag. They are two different shakes wearing the same number, and they should not be swapped for each other inside a week as though they were interchangeable. We did not add a fiber term to the Score. Nothing above earns it a weight.
References
Jenkins DJA, Wolever TMS, Taylor RH, Barker H, Fielden H, Baldwin JM, Bowling AC, Newman HC, Jenkins AL, Goff DV. Glycemic index of foods: a physiological basis for carbohydrate exchange. American Journal of Clinical Nutrition, 34(3):362-366, 1981. | Meng H, Matthan NR, Ausman LM, Lichtenstein AH. Effect of macronutrients and fiber on postprandial glycemic responses and meal glycemic index and glycemic load value determinations. American Journal of Clinical Nutrition, 105(4):842-853, 2017.
Keep reading
Previous experiment
Not sure which shake fits you? We match the formulation to your body type.
Find your body type