Start with the cocoa flavor you want to retain. Worldwide Botanicals gives its two retail powders different roles: the medium roast is positioned for regular cocoa and milk-chocolate-style recipes, while the dark roast is positioned for dark cocoa and dark-chocolate-style recipes. These are the seller’s suggested uses, not results of an equivalence test.[6][7]
That distinction is more useful than looking for a generic “cocoa substitute.” It helps define the first sample comparison. Deciding how much to use, whether to change wheat flour and whether the cake will become denser still requires formulation evidence. Two baking studies illustrate the central issue: similar acceptance does not mean identical flavor, and a published percentage is not necessarily a cocoa replacement rate.[3][4]
Medium roast, dark roast and Raw: three product positions
| Product and ASIN | How the listing positions it | Question for a sample trial |
|---|---|---|
| Product and ASINWorldwide Botanicals medium roast, B07JX6YCKR | How the listing positions itUses “regular cocoa” and recipes calling for milk chocolate as reference points.[6] | Question for a sample trialDo sweetness, cocoa taste and baked aroma work together in the target cake? |
| Product and ASINWorldwide Botanicals dark roast, B0F67QBGL3 | How the listing positions itUses “dark cocoa” and recipes calling for dark chocolate as reference points.[7] | Question for a sample trialDoes the baked product achieve the intended color and flavor, rather than merely starting with a darker powder? |
| Product and ASINThe Australian Carob Co., B01D2975KS | How the listing positions itTitled “Raw,” under The Australian Carob Co.; product bullets are absent from the detail sample.[8] | Question for a sample trialIf the aim is a deliberately carob-flavored product, is this worth testing alongside roasted candidates? |
The medium and dark products belong to the same brand, making them useful examples of how a seller segments its range by target flavor. The Raw product is not that brand’s third roast level. A cross-brand comparison can also introduce differences in raw material and processing, so any observed difference cannot automatically be attributed to roasting alone.[6][7][8] The Raw detail capture also labels the item type “Video Game,” inconsistent with its title. We therefore use only its title and brand as positioning evidence, not as confirmation of processing temperature, nutrition or quality.[8]
A milk-chocolate-style use case does not establish that a powder can replace the milk chocolate in a recipe gram for gram. Nor do these listings define “dark cocoa” as a specified alkalization level or pH.[6][7] A useful sample request should therefore state whether the existing ingredient is cocoa powder or chocolate, and which flavor attributes must be retained—not simply request medium or dark roast.
These pages offer positioning, not a controlled performance comparison. We do not adopt their health, low-calorie, cost-saving or “perfect replacement” claims, or infer sales from search positions and review counts. Roast choices also exist in industrial supply: Nordmann’s introduction to Carob Ingredients lists light, medium and dark roasted pulp powders for baked goods and cocoa drinks, but provides no same-recipe comparison across those roasts.[2]
“30% carob” does not mean replacing 30% of the cocoa
A 2020 Madeira-type cake study used commercial roasted flour from Creta Carob, not the three retail samples above. Its 100% flour basis was the combined 75 g of wheat flour, cocoa and carob flour.[3]
| Study treatment | Wheat flour | Cocoa powder | Carob pulp flour | Carob share of the listed 300 g ingredient total, calculated here |
|---|---|---|---|---|
| Study treatmentControl | Wheat flour60 g | Cocoa powder15 g | Carob pulp flour0 | Carob share of the listed 300 g ingredient total, calculated here0 |
| Study treatment10 | Wheat flour67.5 g | Cocoa powder0 | Carob pulp flour7.5 g | Carob share of the listed 300 g ingredient total, calculated here2.5% |
| Study treatment30 | Wheat flour52.5 g | Cocoa powder0 | Carob pulp flour22.5 g | Carob share of the listed 300 g ingredient total, calculated here7.5% |
| Study treatment50 | Wheat flour37.5 g | Cocoa powder0 | Carob pulp flour37.5 g | Carob share of the listed 300 g ingredient total, calculated here12.5% |
| Study treatment70 | Wheat flour22.5 g | Cocoa powder0 | Carob pulp flour52.5 g | Carob share of the listed 300 g ingredient total, calculated here17.5% |
Each treatment also contained 75 g margarine, 75 g sugar and 75 g egg; these weights come from the paper’s Table 1.[3] Treatment 30 removed all 15 g of cocoa, added 22.5 g of carob and reduced wheat flour from 60 to 52.5 g. Treatment 10 removed the cocoa while increasing wheat flour to 67.5 g. This was not a series that changed only the fraction of cocoa replaced.[3]
The final column is a calculation on the listed ingredient total, not the baked product’s mass. The materials section mentions baking powder, but Table 1 gives no amount, so the table explains the substitution structure rather than providing a complete replication recipe. Figure 1’s caption also labels the control as 15% cocoa on a flour basis, conflicting with Table 1’s 15 g ÷ 75 g = 20% and the 20% stated in the text. The comparison here follows the tabulated weights.[3]
Treatment 30 is a useful reference—not unchanged texture
Twenty trained panelists assessed the cakes 24 hours after baking, with three samples per person and 12 assessments per treatment. The study used a 15 cm line scale. Overall acceptability averaged 10.4 for treatment 30 and 8.9 for the control, but every treatment shares the letter “a” for this attribute in Table 3. The higher mean does not establish a preference advantage.[3]
Physical properties also need to be read separately. Table 2 gives specific volume as 2.22 ± 0.12 mL/g for the control (a), 1.93 ± 0.00 for treatment 30 (abc) and 1.78 ± 0.10 for treatment 50 (bc). Treatment 30 and the control share a letter and are not significantly different; treatment 50 and the control do not, and treatment 50 has significantly lower specific volume. Failure to detect a significant difference is not evidence of equivalence.[3]
Figure 1 adds an important texture distinction. Treatment 50 had a significantly harder crumb than the control, while treatments 10 and 30 were significantly softer. Treatment 30 did not differ significantly from the control in springiness, but its cohesiveness was already significantly lower; this change did not begin only above 30%. Springiness and cohesiveness declined further at treatments 50 and 70.[3]
For a project aiming to retain a relatively soft cake while changing flavor, treatment 30 is a useful trial reference—not an optimum with unchanged texture. The authors also called for a standard acceptance test with 100 consumers to confirm the sensory findings.[3]
The muffin trial: acceptable, but still different
A 2018 muffin study used either 5.0% cocoa or 5.0% carob powder. Here, 5% belongs to the authors’ listed whole-formula proportions, not a 75 g flour basis. The formula also contained 25.5% soybeans, 6.5% flaxseeds and 6.0% wheat flour.[4] The soybeans were soaked for 12 hours, cooked for 45 minutes and blended. They should not be interpreted as 25.5% dry soy flour; the paper does not clearly state whether that proportion was weighed before or after soaking.[4]
In a nine-point hedonic test with 78 volunteers, carob muffins scored 7.1 and cocoa muffins 6.9, with no statistically significant difference.[4] A separate descriptive assessment by ten trained panelists found the carob muffins sweeter, less bitter and weaker in cocoa taste, with stronger beany and “unknown” taste notes.[4] The development opportunity is a different flavor that remains acceptable, not a promise that substitution will go unnoticed.
The muffin scores should not be compared numerically with the cake scores: a nine-point hedonic scale and a 15 cm line scale were used with different participants and recipes.[3][4] Neither study lowered added sugar as part of its carob comparison. The muffin formula retained 19.0% white sugar, and each cake formula retained 75 g sugar. Greater perceived sweetness is not a successful sugar-reduction trial.[3][4]
Locust bean gum addresses another formulation problem
Pulp powder is the relevant candidate for changing cocoa flavor. If the task is thickening or stabilizing an aqueous system, investigate seed-endosperm-derived locust bean gum, or LBG, separately. CAROB S.A. differentiates its PALGUM range by viscosity, particle size and color, and describes synergistic interactions with xanthan and carrageenan.[1]
Request the exact grade and its test conditions. EFSA’s 2017 scientific opinion describes LBG as swelling in cold water and requiring heating for maximum solubility; this does not prescribe one hydration temperature and time for every commercial grade.[5] The food-grade section of the public PALGUM page also does not provide a complete viscosity test protocol.[1]
One further distinction matters: Nordmann’s VISCOGUM CGP is a seed-derived protein flour, described commercially as containing 40–50% protein and promoted for emulsification. LBG-VISCOGUM on the same page is locust bean gum. The shared VISCOGUM name does not make them the same ingredient, and the protein range is not a test report for a purchased batch.[2]
Turn sample selection into a trial that answers a question
Keep the existing formula as a control, then select medium roast, dark roast or a separate Raw candidate according to the desired flavor. In the first comparison, hold other conditions constant and record carob, cocoa and wheat flour in grams. Record both the powder-fraction percentage and the percentage of all weighed ingredients. Compare baked color, sweetness, cocoa taste, specific volume, softness and cohesiveness before deciding whether to change roast or inclusion level.
If sugar reduction is also required, test it in a separate formulation comparison. If aqueous-phase consistency or stability remains a problem, introduce a specified LBG grade rather than assuming that gum will offset every effect of a large powder-blend change. These are proposed development steps, not a validated production formula.
A useful purchasing conversation can start with: “We want to reduce this cocoa powder while retaining this color and texture.” That gives a supplier more to work with than “30% carob.” Third-party products illustrate selection options; actual supply specifications, documentation and samples still require confirmation.
Sampling scope
Product-positioning comparison: eight Amazon.com search records for “carob powder,” including three detail-page samples (27 September 2026). Brand descriptions inform sample selection; no blind tasting or same-formula product comparison was performed.[9]
Sources
[1] gum
[2] commerce
[3] cake
[4] muffin
[5] efsa
[6] Worldwide Botanicals medium roast listing
[7] Worldwide Botanicals dark roast listing
[8] The Australian Carob Co. Raw product detail, 27 September 2026