In this LAB note
Question · Hypothesis · Method · Results · What Failed · What We Learned · What Changed at MOMS
Question
A shopper picks the green shake over the orange one and assumes she picked up more fiber and fewer calories. Across our 201 formulations, does the color of a whole-food shake predict any of its nutrition — or only which plant went into it?
Hypothesis
We expected color to be a weak but real proxy. The green shakes should sit at the fiber-dense, lower-calorie end because their color comes from leaves and crucifers, and the orange and purple shakes should skew sweeter and more carbohydrate-heavy because their color comes from fruit. We did not expect the effect to be large, but we expected it to be measurable and directionally correct — enough that a customer scanning a row of colors would be picking up a faint real signal rather than a false one.
Method
Two published lines framed the test. Schuldt (Health Communication, 2013) ran a randomized experiment in which participants judged a candy bar to be healthier when its calorie label was printed in green rather than red, despite the two labels carrying identical calorie information — color moved the health judgment while the nutrition stayed fixed. Minich (Journal of Nutrition and Metabolism, 2019) argues the opposite-facing point: an "eat by color" approach is worth teaching because color groups plants by phytonutrient class, and the review is explicit that this is a deliberately simplified consumer heuristic built on general patterns rather than a per-food nutritional claim. We note that a 2020 corrigendum was issued on that review for undeclared conflicts of interest; we use it for its taxonomy, not as an outcome study. Internally, we took the Shake Color hex value stored on each of the 201 production recipes, converted it to HSL, and binned it into pigment families: Orange (n = 93), Yellow (36), Red/Pink (24), Green (20), Purple (20), Neutral (7), Blue (1). The 193 recipes in the five main families formed the analysis set. For each family we computed mean and range of fiber, carbohydrate, fat, protein, and kcal, then partitioned the total fiber variance into between-family and within-family components to get the share of fiber that color actually explains. Separately, we read the Fruit & Vegetable and Body slots of every recipe to check which ingredients each color family was actually made of, and ran the inference in both directions: does a color predict the plant, and does a plant predict the color?
Results
The hypothesis was directionally right and practically worthless. Green shakes do carry more fiber than orange ones — by 1.0 gram. Green (n = 20) averages 8.4 g fiber, 373.6 kcal, 33.2 g protein, 38.8 g carbohydrate, 10.5 g fat. Orange (n = 93) averages 7.4 g fiber, 371.2 kcal, 32.9 g protein, 39.7 g carbohydrate, 9.7 g fat. Across all five families the means sit inside a 1.0 g band on fiber (7.4 to 8.4), a 9 kcal band on energy (371 to 380), and a 0.5 g band on protein (32.9 to 33.4). Purple averages 8.2 g fiber and 42.9 g carbohydrate; Yellow is the highest-fat family at 12.6 g; Red/Pink sits at 7.7 g fiber and 42.2 g carbohydrate. Partitioning the variance, color family accounts for 2% of the total fiber variance across the 193 recipes. Ninety-eight percent of it sits inside the families, and the within-family ranges overlap almost completely: Green runs 2.5 to 14.4 g, Orange 2.7 to 13.4 g, Purple 3.6 to 11.3 g. One matched pair makes the point better than the averages do. Both shakes below are built for the same body type. EY-W2-DETOX-009-V1, "Featherweight," renders a pale green (#C7D98A) from Cabbage 45 g + Broccoli 45 g, on Oat Milk 319 ml with Chickpeas 60 g + Egg White 195 g: 353.9 kcal, 31.6 g protein, 42.6 g carbohydrate, 7.0 g fat, 9.4 g fiber. EY-W3-GUT-001-V1, "Gut Harmony," renders a warm peach (#F9C6A8) from Asian Pear 50 g + Blueberry 35 g + Dates 12.5 g over Carrot 45 g, on Oat Milk 97.5 ml + Soy Milk 150 ml with Chickpeas 30 g + Egg White 195 g: 354.5 kcal, 31.7 g protein, 42.2 g carbohydrate, 7.3 g fat, 9.5 g fiber. The two look like different products on a shelf. They differ by 0.6 kcal, 0.1 g protein, 0.4 g carbohydrate, 0.3 g fat, and 0.1 g fiber. What color does track, it tracks perfectly: all 20 purple shakes contain a purple-pigment ingredient — blueberry, beet, cherry, or taro — such as EY-W2-DETOX-005-V1, "Light Horizon" (#CFBBDD), with Blueberry 43.9 g; and all 20 green shakes contain a green vegetable. Color is a 100% reliable readout of which plant species is in the glass and a 2% readout of what the glass contains nutritionally.
What Failed
The "color equals species" rule only runs one direction, and we found that out the hard way. Green predicts a green vegetable in all 20 cases. The reverse fails outright: 16 recipes containing spinach, kale, or broccoli do not render green at all. EY-W3-GUT-007-V1, "Steady Gut," carries Broccoli 100 g + Cabbage 110 g — 210 grams of green vegetable, among the highest vegetable loads in the database — and finishes at #CBC18E, an olive-yellow, because banana and avocado in the body slot mute the chlorophyll. A customer scanning a page for the greenest option would skip it. Our own product cards were quietly doing that to her. The greenest shake in the set is also a counterexample to our hypothesis. EY-W3-GUT-009-V1, "Grounded Gut," is the most saturated green we make (#8FBF7F) and carries 6.8 g fiber — below the orange family average of 7.4 g — because its color comes from Avocado 45 g + Banana 80 g with only Cabbage 25 g of actual vegetable. Two limits on our own evidence are worth stating plainly. The families are badly unbalanced (Orange n = 93 against Green n = 20, Blue n = 1), so the small-family means are noisy and we would not defend a 1.0 g gap as a real effect. And the hex values are a designer's rendering of the finished blend, not a spectrophotometer reading of it — they are how we depict the shake, which is exactly the signal a customer sees, but it is not an instrument measurement.
What We Learned
Color in this database is a species label, not a nutrition label. It tells you which pigment family went into the glass — chlorophyll, carotenoid, anthocyanin, betalain — and that is a genuinely useful thing to know, because it is the variable Minich's "eat by color" framework is built on and the reason rotating colors across a month is worth doing. It does not tell you fiber, protein, or calories. Reading it that way is the Schuldt effect operating on a smoothie instead of a candy wrapper: the color moves the judgment while the numbers stay where they were. The honest instruction is narrow. Use color to diversify what you eat. Use the numbers to decide what you eat.
What Changed at MOMS
We stopped letting the swatch stand in for the nutrition summary. Color is now handled as a rotation variable rather than a quality signal: when we assemble a member's four-week set we check that it spans at least three pigment families instead of three shades of the same one, and fiber, protein, and calorie targets are set independently and never inferred from the color. On the product side, any shake carrying more than 150 grams of vegetable now states that number on its card even when the blend renders yellow or brown, so a formulation like "Steady Gut" is not passed over for looking less green than it is. We are also no longer treating a deep green swatch as a formulation win on its own — "Grounded Gut" taught us that a green can come from avocado and banana as easily as from a crucifer.
References
Schuldt JP. "Does Green Mean Healthy? Nutrition Label Color Affects Perceptions of Healthfulness." Health Communication, 28(8), 814-821, 2013. | Minich DM. "A Review of the Science of Colorful, Plant-Based Food and Practical Strategies for 'Eating the Rainbow'." Journal of Nutrition and Metabolism, 2019, Article 2125070, 2019.
Keep reading
Not sure which shake fits you? We match the formulation to your body type.
Find your body type