A pilot batch that comes out right on the first attempt is the most misleading result in development. The trial was made in a vessel the production plant does not run, heated through a jacket sized for a fraction of the volume, mixed by an impeller that will never turn in the full-size tank, & cooled in a fraction of the time a production batch needs. The sample a brand signs off is genuine, but it was made under conditions that will not exist again.
This article sets out where the gaps open between a pilot batch & a production run: mixing geometry, heat transfer, the hours a bulk spends waiting, how powders hydrate at volume, what happens to air & volatiles, where a sample is drawn from, & what the filling line does to a bulk that behaved perfectly in the tank. It is written for the buyer who approves the golden sample & then has to explain why the first production run looks different. The route from brief to golden sample is set out in our sampling process guide, & the testing that follows a change in scale is covered in our stability testing guide.
Why a pilot result is not a promise
A pilot answers whether a formula is viable. It does not answer whether a process is capable. Viability is a property of the composition: the emulsifier system holds, the preservative passes, the texture is what the brand asked for, the fragrance survives the base. Capability is a property of the plant, & the plant that made the pilot is rarely the plant that will make the order.
Three variables change with volume & cannot be reproduced by writing a bigger number into the same instruction. The first is geometry: the ratio of impeller diameter to tank diameter, & the ratio of liquid height to tank diameter, both of which shift when the vessel changes. The second is thermal history: how fast the mass heats, how long it holds hot, & how fast it cools. The third is time in the sense of how long each unit of the bulk spends under shear or under heat, which is what actually decides droplet size, hydration & structure.
The practical consequence is that a pilot should be read as a formula check, then confirmed by a scale-up trial in the vessel that will run the order. A brand that skips the scale-up trial is not saving a step. It is moving the discovery out of a controlled trial & into a paid production run with a launch date attached to it.
Mixing geometry, tip speed & power per volume
Two numbers describe a mixer, & they pull in opposite directions when the batch grows. Tip speed is how fast the outer edge of the impeller moves through the liquid, which governs the shear that breaks droplets & wets out powder. Power per unit volume is how much energy goes into each litre, which governs turnover, suspension & how quickly an additive distributes through the whole mass.
In a small vessel both are easy to satisfy. In a large one, holding tip speed constant while the volume rises demands a disproportionate increase in power, so plants typically hold power per volume & accept a lower local shear, or hold tip speed & accept slower bulk turnover. Either choice makes the bulk behave differently from the pilot, & the difference shows up as texture, viscosity & dispersion quality rather than as an obvious failure.
The instruction that fails most often is the mixing time copied straight from the pilot record. Ten minutes in a small vessel may equal many tank turnovers; the same ten minutes in a production vessel may be two. The fix is to write the end point rather than the clock: a specified number of turnovers, or a stated appearance such as full incorporation with no visible powder, alongside the time as a secondary guide.
A brand should ask which impeller runs the vessel, its diameter relative to the tank, the speed in revolutions per minute for each stage, & the fill level the vessel was designed for. A vessel run well below its design fill level does not mix the way the drawing suggests, & that is worth knowing before the first production slot is booked.
Heating, cooling & the hours the bulk waits
Surface area per unit volume falls as a batch grows, so a large mass heats more slowly & cools more slowly than a small one even with the same jacket temperature. A pilot that reaches temperature in minutes can correspond to a production run that takes a substantial part of a shift, & the bulk is not idle during that time. It is hot, it is being stirred, & it is losing volatile material.
The hot hold is where formulation decisions get tested. Waxes & crystalline structurants melt & then set according to the cooling curve, & a slow cool lets crystals grow into a different network from the one the pilot produced. The result is a cream that is firmer or softer than the approved sample with no change in the formula at all. Emulsification temperature behaves the same way: a longer approach to the target changes the droplet size before the homogeniser ever runs.
Cooling is also the stage most often left undescribed. A pilot record that says cool to a set temperature says nothing about how fast, yet the rate decides structure, viscosity build & the point at which heat-sensitive materials can be added. The production specification should state the cooling profile, not only the end temperature.
Powder hydration & dispersion at volume
Polymers, clays & pigments are the materials most sensitive to how they enter the water. Added to the surface of a vortex, a hydrating polymer forms a gel layer around each particle, & the clumps that follow are slow to break down. In a beaker a spatula & patience solve it; at production volume the same behaviour produces a batch with specks, a viscosity below target, or a texture that keeps firming for days after filling.
The scale effect is not simply that there is more powder. The vortex in a large vessel is broader & slower to refresh, the powder goes in over a longer period, & the material charged first has been hydrating while the rest is still being added. An instruction that reads add over five minutes has a different meaning at each scale.
The controls are well known & worth naming in the brief. A pre-slurry in a non-solvent such as glycerin or a glycol keeps particles separated until they reach the water. An eductor or inline disperser draws powder straight into a high-shear zone instead of onto a surface. A side pot for the gum phase, hydrated separately & then added, removes the problem from the main vessel entirely.
The factory should state which of these it uses, & the batch record should show the hydration time before the next stage begins. Where a batch is released on a viscosity measured too early, the number reflects incomplete hydration rather than the formula, & the drift continues in the warehouse.
Droplet size, homogenisation & emulsion texture
The pilot homogeniser is usually a batch rotor-stator head immersed in a small vessel. Production may use an inline head through which the bulk passes once per recirculation, or a high-pressure unit with a fixed gap. The shear mechanism, the residence time in the shear zone & the temperature at that moment all differ, so the same name on an equipment list does not guarantee the same droplet size.
The quantity that matters is how many times the whole batch passes through the shear zone, & how long it spends there per pass. A small batch recirculates through a head many times over in a short period; a large batch may pass once at a flow rate chosen to finish the transfer quickly. Matching the result means matching the passes & the temperature, not matching the machine name.
Over-processing is the opposite failure. Extended recirculation can break a polymer-thickened structure, drive air in, or warm the bulk past the point the specification allows. A factory that reports homogeniser minutes without reporting temperature & pass count has given a number that cannot be compared with the pilot.
Air, foam & de-aeration
Small batches release entrained air quickly because the distance to the surface is short & the headspace is close. A tall production vessel holds a foam blanket that can persist, insulates the surface, slows cooling, distorts level readings & changes what the filler draws. Air that the pilot never showed becomes a production characteristic.
The consequences are practical rather than cosmetic alone. Entrained air lowers the apparent fill weight, so a pack filled by volume can fall under the declared quantity. Bubbles in a clear gel or a serum are visible on the shelf. Dissolved oxygen in the bulk is a variable in oxidation of oils & in colour drift across the shelf life.
Vacuum de-aeration is the usual answer, & it takes longer at scale, particularly with a viscous or surfactant-rich bulk that holds foam. Adding an antifoam is faster but introduces a material that affects clarity, foam performance in the wash & the ingredient declaration. A brand should know which route the factory intends before the run, not after a fill weight check fails.
Volatile loss: fragrance, alcohol & water
Everything that heats a large mass for a long time drives off the lightest fractions first. A fragrance balanced in a small batch can lose its top notes over a long hot hold, leaving the production bulk smelling flatter or more sharply of its base. The formula did not change; the composition at the end of the run did.
Water behaves the same way, & the effect is easy to miss because the loss is small in percentage terms & large in consequence. A bulk that loses water becomes more concentrated, which raises viscosity, shifts pH & changes the effective concentration of the preservation system. Alcohol-containing systems lose strength in the same way, & the label declaration was written for the formula rather than for what remained after the hold.
The control is straightforward where the factory is willing to document it: a make-up charge of water or alcohol at the end of the process, recorded with the quantity, the measurement behind it & the reason. Where a condenser or reflux arrangement is fitted, the factory should say so, because it changes what the pilot needs to predict.
A brand should also expect the fragrance dose to be reviewed after the first production run. The level that was right in a vessel with a short hot hold can read weak in a bulk that spent hours at temperature, & adjusting it after the run is cheaper than shipping a product whose scent no longer matches the approved sample.
Sampling points & what a sample can hide
A sample is only as representative as the place it came from. A valve at the bottom of a vessel draws material that has been sitting under the impeller; a dip from the top draws whatever is floating. In a pigmented or suspended system the two are not the same product, & neither is necessarily what the filler sees.
Recirculation before sampling is the standard control, & the duration matters. Five minutes of recirculation in a pilot vessel homogenises the bulk; the same five minutes in a production vessel may not. The specification should state the recirculation period before a sample is drawn, & the batch record should show it was observed.
The sample that predicts the customer's experience is the one drawn at the filler, after the line is flushed & before the first saleable pack, plus packs retained from the start, middle & end of the run. A tank sample approved at the start of filling tells a brand about the tank, not about the pack, & the pack is what is sold.
Line flush is the related risk. Water or rinse left in the pipe dilutes the first packs, which are then either scrapped or, worse, released. The factory should state how much product is run to waste before filling begins, & that quantity should be agreed rather than left to the operator.
Transfer, filtration & the filling line
Between the vessel & the pack the bulk meets a pump, a filter & a length of pipe, & each one changes it. A lobe or gear pump shears a structured cream; a peristaltic pump pulses; a filter removes oversized particles but can also strip the structure that gives the product its body. The bulk arriving at the nozzle is not the bulk that was approved in the tank.
Line hold-up is a yield question that becomes a cost question on short runs. Several kilograms of product stays in the pipe & the hopper, & on a small run that is a meaningful share of the batch. For a brand with a tight budget this is worth asking about at quotation stage, because it changes how much bulk has to be made to deliver a given number of packs.
Temperature drift during filling is the quiet one. A bulk that cools in the hopper thickens, & a volumetric filler then delivers less mass per pack; a bulk that warms thins, drips or over-fills. Settling matters in the same way for pigmented & suspended systems, which is why hopper agitation & fill weight checks at intervals through the run belong on the record rather than in an optional extra column.
What the pilot batch record should capture
The purpose of a pilot record is to be reproducible by someone who did not attend the trial. That means charge weights, the order of addition, the temperature at each stage, mixer speed & duration for each stage, homogeniser settings with pass count & temperature, vacuum time, pH, viscosity with the spindle & the temperature at which it was read, appearance, odour & yield.
Viscosity is the line most often written without its conditions. A figure read on a small sample at an unstated temperature with an unstated spindle cannot be compared with a production result, & the argument that follows is unresolvable. Insist on the method being named on the specification sheet, not only the value.
Deviations deserve a line of their own. Extra water added to hit a viscosity, a re-mix, a re-heat, a second pass through the homogeniser: each one is a change to the process, & each one should appear on the record with the reason. A pilot record with no deviations is either a very good trial or an incomplete one, & the brand should ask which.
Finally, the golden sample has a shelf life. A physical reference jar changes colour, texture & odour over months, so a written specification plus a freshly made reference from each production run is more reliable than a jar kept on a shelf. Where the two disagree, the written specification should govern, & the brand should confirm which document the factory works to as the standard.
Scaling checks at a glance
The table below sets the checks that matter most against what changes when a batch grows, what that does to the bulk, & what a brand should ask the factory to state in writing. The rows are a working map rather than a fixed rule; the size of each effect depends on the formula, the vessel & the plant, & should be discussed before the first production slot is booked.
| Check | What changes at scale | What it does to the bulk | What to ask for |
|---|---|---|---|
| Vessel geometry | Impeller fit & liquid height shift | Different turnover & dispersion | Impeller type, speed range, fill level |
| Mixing time | Same minutes give fewer turnovers | Under-mixed or over-sheared bulk | A turnover criterion plus a visual end point |
| Heating rate | Surface area per volume falls | A longer hot hold for the mass | A time & temperature log through heat-up |
| Cooling profile | A large mass cools slowly | Different crystal structure & viscosity | The cooling curve stated on the record |
| Powder addition | More powder into a broader vortex | Lumps, slow hydration, air pick-up | Pre-slurry, eductor or side pot method |
| Homogenisation | Passes per unit volume differ | A shift in droplet size & texture | Head type, pass count, stage temperature |
| De-aeration | A foam blanket holds in a tall vessel | Bubbles, low fill weight, oxidation | Vacuum hold time & foam handling |
| Transfer & filling | Line hold-up, shear, hopper cooling | Yield loss & fill weight drift | First-off checks & retained packs |
The rows are a starting map, & the right set is matched to the formula & the plant rather than copied from a generic sheet. A brand that asks for the geometry, the thermal profile, the pass count & the sampling points gets a production run it can compare with the pilot. A brand that asks only for the golden sample to be matched gets an outcome it cannot explain when the two differ.
What to put in the brief
Scale-up questions cost nothing at brief stage & are expensive once the run is scheduled. State the volume of the first production run, because the vessel that suits it may not be the vessel that made the pilot, & the brand should know which one will be used. State what the launch quantity is & how soon a repeat is expected, since a repeat at a different volume repeats the exercise.
- The pilot size & the intended production size, & whether a trial in the production vessel is included.
- The vessel that will run the order, with the impeller type, speed range & design fill level.
- The heating, holding & cooling profile, written as time & temperature rather than as set points alone.
- How powders & gums are incorporated: pre-slurry, eductor or side pot.
- The homogeniser type, the pass count or recirculation time, & the temperature at that stage.
- The de-aeration method, including the vacuum hold & how the foam layer is handled.
- Sampling points & recirculation time before sampling, plus packs retained from start, middle & end.
- The viscosity method with spindle & temperature, & the fill weight check frequency during the run.
Then agree what happens when the first production run does not match the pilot. A brand that settles this in advance gets a documented investigation; a brand that raises it after delivery gets an argument about whose sample is the reference.
Where our capability fits
We run pilot batches & scale-up trials in the same plant that runs production, & the trial is written up as a batch record rather than as a note. The record carries the thermal profile, the mixing stages, the homogeniser settings & the sampling points, so the production instruction is written from data instead of from memory. Where a formula needs a vessel we do not run, we say so before the trial rather than after it.
We work with brands on private label (your brand on a proven base), OEM/ODM contract manufacturing & full custom formula development, & the scale-up discipline is the same across all three. Our lines include a private label moisturising shower gel, a hydrating peptide cream jar & a hyaluronic acid hydrating serum, each made under the same trial-to-production routine.
Tell us the formula, the pilot size & the volume you plan to launch at, & we will come back with which vessel would run it, what the trial would measure & where we expect the gaps. Start at our inquiry page, email capsulebiotech@163.com, or message the right team on WhatsApp: Capsule Series (capsules only) +86 158-7655-5319, Cosmetics Series (serum, hair care, body care, fragrance) +86 158-7655-5319.
General manufacturing information, not regulatory or legal advice. Scale-up behaviour depends on the formula, the vessel & the plant, & the specification, testing & documentation that apply to a given product should be confirmed with the factory & a qualified professional before launch.