Food Product Development, Formulation & Labelling

From idea to supermarket shelf. Learn the stage-gate product development process, consumer research and sensory evaluation (difference, descriptive and hedonic tests), formulation and standardisation (Pearson square, recipe scaling), ingredient functionality (emulsifiers, hydrocolloids, sweeteners), nutrition labelling and claims, costing and pricing, scale-up and pilot trials, plant layout and regulatory approval, plus trends like plant-based, fortified and clean-label foods — with Python for milk standardisation and a per-serving nutrition label.

Most new food products fail in their first year. Successful development combines consumer insight, food science, engineering, regulation and cost control through a structured process.

Food technologists formulate recipes, test sensory appeal, ensure safety and shelf life, calculate nutrition labels, and scale up from kitchen to factory.

An analogy

Developing a food product is like writing and producing a hit song: an idea that resonates (consumer need), skilled composition (formulation), rehearsals (pilot trials), audience testing (sensory panels) and a production team that can reproduce it perfectly every night (manufacturing).

Stage-gate development

Typical stages:

• Idea generation — trends, consumer insights, technology push
• Concept screening and business case
• Formulation and prototype development (bench/kitchen)
• Sensory testing and consumer trials; shelf-life and safety studies
• Scale-up, pilot plant, packaging, regulatory approval, launch and review

Formulation and ingredient functionality

Ingredients have functions beyond nutrition: emulsifiers (lecithin, mono/diglycerides) stabilise oil–water mixtures; hydrocolloids (guar, xanthan, pectin, agar) thicken and gel; starches give body; sweeteners (sucrose, sugar alcohols, stevia) affect taste, bulk and browning; salt affects flavour and preservation; leavening agents, acids and enzymes shape texture. Changing one ingredient (e.g. sugar reduction) often affects many properties.

Sensory evaluation

Difference tests (triangle, duo-trio) check whether a change is noticeable; descriptive analysis with trained panels profiles attributes; affective tests (9-point hedonic, just-about-right) measure consumer liking. Controlled booths, coded samples, randomisation and statistics (ANOVA) ensure valid results.

Labelling, cost and scale-up

Labels include name, ingredients (descending order), allergens, nutrition per 100 g/serving, veg/non-veg logo (India), FSSAI licence, batch, dates, storage and claims (which must meet regulatory criteria). Costing adds ingredients, packaging, processing, overheads and margin. Scale-up issues: mixing times, heat transfer, ingredient variability, equipment differences. Trends: plant-based dairy/meat, millets, protein fortification, sugar/salt reduction, upcycled ingredients, personalised nutrition.

Formulas

• Pearson square: parts A = |target − B|; parts B = |A − target|
• Per-serving value: value/serving = value/100 g × serving (g)/100
• % energy from fat: 9 × fat/energy × 100
• %DV: nutrient per serving/reference value × 100
• Selling price: price = COGS/(1 − margin)

Worked example: Pearson square

Given: Mix 40% and 10% sugar syrups to 25%

Find: Parts.

• 40%: 25 − 10 = 15 parts
• 10%: 40 − 25 = 15 parts

Answer: 1 : 1

Worked example: Per serving

Given: Sodium 650 mg/100 g, serving 30 g

Find: Sodium per serving.

• 650 × 0.3

Answer: 195 mg

Worked example: Selling price

Given: COGS ₹24, desired margin 40%

Find: Price.

• 24/0.6

Answer: ₹40

Milk standardisation and nutrition label

# Pearson square: blend milks to standardise fat, and a nutrition label per serving
fat_a, fat_b, target = 6.0, 0.1, 3.0          # % fat: full-cream, skim, toned-milk target
parts_a, parts_b = target - fat_b, fat_a - target
batch = 1000.0
kg_a = batch * parts_a / (parts_a + parts_b)
print(f"full-cream {kg_a:.1f} kg + skim {batch-kg_a:.1f} kg -> {target}% fat")
# Nutrition label for a 30 g serving of a snack (per 100 g values)
per100 = {"energy kcal": 520, "protein g": 7.0, "fat g": 30.0, "sat fat g": 12.0, "carbs g": 55.0, "sugar g": 4.0, "sodium mg": 650}
for k, v in per100.items():
    print(f"{k:12s} per 100 g {v:6.1f} | per serving {v*0.30:6.1f}")
print(f"energy from fat {per100['fat g']*9/per100['energy kcal']*100:.0f}%")
# full-cream 491.5 kg + skim 508.5 kg -> 3.0% fat
# energy kcal  per 100 g  520.0 | per serving  156.0
# protein g    per 100 g    7.0 | per serving    2.1
# fat g        per 100 g   30.0 | per serving    9.0
# sat fat g    per 100 g   12.0 | per serving    3.6
# carbs g      per 100 g   55.0 | per serving   16.5
# sugar g      per 100 g    4.0 | per serving    1.2
# sodium mg    per 100 g  650.0 | per serving  195.0
# energy from fat 52%

The Pearson square quickly finds blend ratios: nearly equal parts of 6% and 0.1% milk make 3% toned milk. Labels declare nutrients per 100 g and per serving; this fried snack gets 52% of energy from fat, a target for reformulation (baking, air-popping or reduced-fat recipes).

Common mistakes

• Wrong: Launching without consumer testing. — Right: Use sensory and market tests early.
• Wrong: Making claims without meeting criteria. — Right: Check FSSAI claims regulations.
• Wrong: Assuming bench recipes scale linearly. — Right: Pilot trials reveal mixing and heating differences.

Applications

• Dairy standardisation: Toned and double-toned milk.
• Healthier reformulation: Salt, sugar and fat reduction.
• Millet products: Promoting nutri-cereals.
• Plant-based foods: Dairy and meat alternatives.

Key takeaways

• Stage-gate: idea → prototype → test → scale → launch.
• Pearson square blends to targets.
• Label per 100 g and per serving; check claims.
• Ingredients have multiple functions — reformulate carefully.

Going further

• Mixture design (DoE): Optimising multi-ingredient formulations.
• Sensory statistics: Preference mapping and ANOVA.
• Precision fermentation: Animal-free proteins.