Why a flavour that works in water collapses in milk

A flavour is dosed, approved in a water-based trial, and then goes into a milk drink and arrives faint and muddled. Nothing was measured wrong.

The cause is partitioning. Aroma compounds are mostly lipophilic — they dissolve happily in fat. In a clear beverage there is almost nowhere for them to go except the headspace above the liquid and, eventually, your nose. Add fat and you have given them somewhere far more comfortable to sit. They stay dissolved in the fat phase, less reaches the headspace, and the product reads weaker.

Protein compounds the effect. Milk proteins bind some aroma compounds directly, so a portion of what you dosed is effectively taken out of circulation.

Why more flavour is the wrong answer

Fat does not suppress every compound equally. The more lipophilic ones are held back hardest, while the lighter, more polar top notes carry on much as before. So the profile does not just get quieter — it changes shape. Doubling the dose brings back the top notes long before it recovers the body, and you end up with something sharp and thin at a cost you did not budget for.

The profile has to be rebalanced for the base it is going into. That is why a brief that begins "we already have this flavour, we just want it in a yoghurt" is a development job and not a re-dose.

The same logic runs the other way. Fat is also what makes a dairy flavour last — it releases slowly through the whole mouthful instead of arriving and vanishing. It is why full-fat ice cream tastes rounder than a sorbet made with the same flavour, and part of why reduced-fat products are hard to get right.