The life of a flavour begins not in the laboratory but at the briefing table. A sentence left unsaid there will come back to meet you on the production line three months later.
When a customer calls and says "we need something with vanilla", there is not yet a request on the table. There is an intention. Turning that intention into a workable technical definition is the most critical and the most underrated step in flavour development. Everything that follows — trials, samples, revisions, scale-up — depends on the quality of that first step.
A good brief names the flavour; what really determines the outcome is that it describes the world the flavour will have to live in. The same "strawberry" request produces three unrelated formulations in a hot-fill beverage, in an extruded snack and in a moist shisha tobacco blend. The matrix of the product, the temperature of the process, the target dosage, the market it is going to and the time it will spend on the shelf — each of these pulls the recipe in a different direction from the outset. When that frame is not put on the table, the laboratory can start working, but it does not know what it is working against; and in a project, time is usually lost not at the bench but in the unnecessary rounds run because the brief was never sharpened.
The first direction: several possibilities at once
Once the brief is settled, the flavourist turns to the archive. This is where Radiance's library of more than 500 formulations comes into play: instead of starting from zero, the starting points closest to the target are selected. If there is a reference product, its profile is mapped with analytical methods such as GC-MS; the resulting list is a map showing where the work should begin.
In the first round we prepare not a single sample but usually two to four different directions. One aims at the exact centre of the brief, the others deliberately veer to either side: one fresher and greener, one riper and more cooked, one more persistent. The goal is not to hit the target on the first attempt but to open a space of choice in which the customer can say "not this one, but go in that direction". If you send a single sample, the feedback you get will be equally one-dimensional.
Revision rounds: putting feedback through translation
Comments from a customer rarely arrive in technical language. They are sentences such as "it feels a bit artificial", "the finish is too short", "it does not hit my nose". The real craft of flavour development lies in translating those sentences into decisions about components and ratios. "Short finish" usually means the volatile and middle notes dominate while the persistent body stays weak. "Artificial" generally means one characteristic component is too dominant and the background is not supporting it.
In these rounds our own sensory panel ties the customer's subjective comment to a repeatable measurement. The panel does not collect preferences, it describes: intensity, character, finish duration and side notes are each scored separately. Preference is one person's opinion; a profile is the team's shared record. It is the second that drives the revision.
“A flavourist's job is not to create a taste but to translate the sentence in someone's mouth into a ratio in the formulation.”
Two or three rounds are usually enough on a well-defined project. If the rounds run past five or six, the problem is not the flavour; there is an ambiguity in the brief, or in who the decision-maker is. In that case, returning to the table gets you there faster than continuing to produce samples.
Application testing: in the real product, in the real process
A flavour almost always looks cleaner in a neutral solution than it really is. The true test begins when it enters the target matrix. Fat holds volatile components and slows their release; sugar softens perception; protein binds certain notes; an acidic environment transforms some components over time. That is why every approved sample is retried in the customer's own product, with the customer's own process.
In heat-treated products, volatile loss is a serious matter, and the solution is often not to strengthen the formulation but to change the carrier. Encapsulation — powder forms commonly produced by spray-dry — locks the flavour inside a carrier matrix: the components pass the hottest point of the process behind a wall, and are released at the consumer end when moisture and chewing dissolve the matrix. The choice between a liquid flavour and a powder flavour is not aesthetic; it is an engineering decision.
Stability work starts at the same stage. The product is monitored under real and accelerated conditions across its target shelf life and comes back to the panel at defined intervals. It is not enough for a flavour to be perfect on day one; it has to still be itself in the sixth month. Colour drift, loss of top notes, off-tastes formed by oxygen passing through the packaging — these are caught at this stage, not in the field.
Scale-up and the first production batch
A formulation that works on 100 grams in the laboratory does not behave identically at tonnage. Mixing time, temperature, order of addition, even the cleaning cycle of the line all affect the result. Scaling up requires rewriting the process rather than multiplying the recipe by a factor. That is why the first production batch is treated like a new product: it is compared analytically and sensorially against the approved sample as reference.
What comes out of that comparison is the anchor against which every subsequent batch will be measured. The gold standard sample, the acceptance ranges and the quality control criteria are all put on record. A year later, when a batch produced from a different raw material lot is under discussion, this is where everyone looks. Without a record, the discussion locks up on the sentence "I do not think this is the same as it used to be", and there is no way out of it.
What actually speeds the process up
What really accelerates a project is not the pace of the laboratory but when the decisions are taken. A brief that is settled early, structured feedback and a production team involved in the application testing from the start will shorten the schedule round by round.
In practice this comes down to two simple decisions. First: the person who will give the final approval should be at the table for the first briefing meeting. A decision-maker who joins later restarts rounds that were already complete. Second: do not leave application testing to the end. Try the first-round samples on your own line, with your own process; an impression taken in a neutral solution is no substitute for one taken in the real product. In most projects these two habits erase two revision rounds outright.
