A botanical that makes an excellent tea bag can make a poor beverage. The difference is not quality in any general sense — it is that a tea bag is brewed once, by a consumer, for three minutes, in a mug, and then thrown away, while a beverage base is extracted at industrial scale, filtered, standardised, blended, packed and expected to taste identical eighteen months later in a can on a warm shelf. Those two jobs place almost opposite demands on particle size, microbiology and lot-to-lot consistency.
This guide is written for the people who feel that gap first: beverage formulators and technical buyers at RTD, cold brew, kombucha, infusion-base, syrup and functional-drink brands who are specifying dried tea and herbal material from a botanical supplier for the first time, or who have inherited a specification that keeps producing surprises at the extraction kettle. If your product is a consumer tea bag rather than a drink, our herbal tea blend development guide is the more useful starting point.
Why beverage applications change the specification
Four things happen when a botanical goes into a beverage process rather than an infusion bag.
Extraction becomes a controlled unit operation. Water temperature, time, agitation, water-to-solids ratio and pH are all set by you, repeatedly, in a vessel. Small differences in the raw material's surface area therefore show up as measurable differences in yield and in sensory profile — not as the variability a consumer would never notice.
Filtration becomes a bottleneck. Every gram of fines and dust that a tea bag would have retained behind filter paper is, in your process, a particle that must be separated. Fines blind filters, extend cycle times, increase losses and cloud a product that was meant to be bright.
The product is stored, not consumed immediately. A pasteurised or hot-filled beverage has a shelf life to defend against microbial and oxidative change. Raw material microbiology and moisture become process-safety inputs rather than background information.
Sensory tolerance narrows dramatically. A consumer who buys the same tea twice accepts a noticeable difference between crop years. A consumer who buys the same canned drink twice does not. Your supplier's lot-to-lot variability becomes your product's consistency problem.
Cut size: the single most under-specified parameter
Most first specifications say "dried chamomile flowers, food grade" and stop. The result is that each shipment arrives at whatever cut the supplier's last customer wanted. Cut size deserves a numbered line in the specification, because it drives extraction rate, filterability and dosing accuracy at once.
The general physics is straightforward: smaller particles have more surface area per gram, so they extract faster and more completely, and they also produce more fines, block filters, and can pull through bitter or astringent fractions that a coarser cut would have left behind. There is no universally correct cut — there is a correct cut for your equipment and your target profile.
| Application | Typical cut character | Why it fits | Main risk |
|---|---|---|---|
| Hot extraction for RTD base, kettle or percolation | Medium cut, tea-bag-cut range, low dust | Fast, near-complete extraction within a normal hold time | Over-extraction of tannins; filter loading if fines are uncontrolled |
| Cold brew and cold infusion | Coarse to medium cut, whole leaf or flower for delicate botanicals | Long contact time compensates for lower surface area; cleaner, sweeter profile | Under-extraction if the cut is too coarse for the hold time; higher dosage needed |
| Continuous or in-line extraction | Uniform medium cut, tightly controlled fines | Consistent bed permeability; predictable flow | Channelling if particle distribution is wide |
| Concentrate and syrup bases | Medium to fine cut | Maximum yield per kilogram | Turbidity and sediment carrying through to the finished product |
| Bag-in-cup and pyramid formats for foodservice | Coarse cut or whole leaf | Visual appeal through the mesh; slow, gentle release | Slow infusion in short service windows |
| Powdered or instant applications | Milled to a defined mesh | Full dissolution or dispersion | Hygroscopic behaviour, caking, dosing accuracy |
What to write into the specification instead of "cut":
- A particle size distribution, not a single number — for example, a stated percentage passing and retained on named sieve sizes.
- A maximum fines or dust fraction, defined by the sieve that separates it.
- A maximum stalk or stem content for leaf and flower material, since stem extracts differently from leaf and changes both yield and taste.
- Bulk density, because your dosing is usually volumetric somewhere in the process even if your recipe is by weight.
- A retained reference sample that both sides agree represents the cut. A physical golden sample settles arguments that written specifications cannot.
Ask the supplier what cutting equipment they run and whether the cut is sieved after cutting. Cutting without downstream sieving means the fines stay in the bag, and the fines are exactly what your filter will find.
Microbiological expectations
Dried botanicals are agricultural products that have been harvested, dried and handled in open air. They carry a natural microbial load. That is normal, and the question for a beverage brand is not "is it sterile" but "is the load compatible with my process".
The parameters that appear on a serious specification are total aerobic mesophilic count, yeasts and moulds, Enterobacteriaceae, E. coli, and absence of Salmonella in a stated sample mass. Bacillus species and their spores matter for any product with a warm hold step, because spores survive pasteurisation and germinate later. Our botanical microbial limits guide sets out how these parameters are tested and what drives them upward.
Your process determines what you can accept:
- Hot-filled or pasteurised RTD. A validated thermal step handles vegetative organisms, so the raw material's total count is less critical than its spore load and its yeast and mould count, which correlate with drying and storage quality.
- Cold brew with a thermal step downstream. Similar, but the cold extraction stage itself is a growth opportunity: cold does not mean microbiologically inert over a long hold, and extraction vessels are nutrient-rich environments.
- Cold brew without a kill step, or fresh chilled formats. The raw material's load largely becomes the finished product's starting load. This is the application where treated material is most often necessary.
- Fermented formats such as kombucha. You want a low and predictable background load, because unpredictable wild organisms compete with your culture and produce off-notes you cannot reproduce or correct.
Where reduction is needed, the common industrial routes are saturated steam treatment and, in some markets, irradiation. Both carry consequences worth negotiating before you commit: steam treatment adds moisture that must be removed again, and both routes can flatten volatile aroma. For delicate florals the sensory cost is real. If you specify treated material, specify it as treated and taste it against untreated before you lock the recipe — the correct decision is a formulation decision, not a purchasing default.
Two other conditions belong in the same section: water activity and moisture content. These govern whether the material stays stable during your storage and whether mould can develop in a warehouse. A botanical delivered at correct moisture and stored badly becomes a different material within a season.
Contaminants, residues and the boring paperwork that stops launches
Herbal infusions attract more contaminant scrutiny than most food categories, because the plant material is extracted rather than eaten, because many species are wild-harvested, and because a few plant families concentrate specific compounds.
The parameters to write into a beverage-grade specification:
- Pesticide residues against the applicable EU maximum residue levels, with a stated scan scope — a multiresidue panel with a named number of analytes is a different assurance from three targeted tests. Our pesticide residue management guide explains how the sampling and the scan scope interact.
- Heavy metals, particularly lead and cadmium, which certain species take up from soil more readily than others. See the heavy metals in botanicals guide for how limits and testing work in practice.
- Pyrrolizidine alkaloids. These plant toxins arise mainly from co-harvested weed species contaminating a crop rather than from the crop itself, and maximum levels for herbal infusions are set in EU contaminants legislation, currently consolidated in Regulation (EU) 2023/915. For anyone buying herbal infusion material this is a first-tier question, and the answer lies in field weed control and harvest discipline, not in downstream cleaning.
- Foreign matter and extraneous plant material, with a stated maximum and a definition of what counts.
- Botanical identity. Species confirmation matters more than buyers expect, especially for genera with close relatives and for material bought through traders. Macroscopic and organoleptic identification is standard; instrumental confirmation is available where the risk justifies it.
On the labelling side, a beverage sold in the EU carries the obligations of Regulation (EU) No 1169/2011 for its ingredient list, allergen declaration, quantity and lot marking. Separately, any suggestion on your pack that an ingredient does something for the body falls under the nutrition and health claims regime of Regulation (EC) No 1924/2006, and only authorised claims may be made. Botanicals are a particularly constrained area here. Design the pack around what you are permitted to say, and treat sensory language as your creative space instead — this is a legal question to settle before artwork, not after.
Batch consistency: what actually causes drift
Consistency complaints almost always trace back to one of five causes, and only one of them is supplier carelessness.
- Crop year. Rainfall, temperature and harvest timing change essential oil content, colour and extract yield. This is agriculture; it cannot be eliminated, only managed. Managing it means blending across lots to a target, holding inventory across a crop transition, and agreeing tolerance bands instead of single-point values.
- Harvest window within a season. Early and late harvests of the same field differ. A supplier who tells you which harvest window a lot came from is giving you a real diagnostic tool.
- Growing region and altitude. Material of the same species from two regions can be reliably different. If a formula depends on a regional character, name the region in the specification and accept that supply is narrower as a result.
- Cut and fines variation. Discussed above, and the most common invisible cause. A lot that extracts faster than the last one is often not a different plant; it is a different particle size distribution.
- Storage and age. Volatiles decline. A lot that has sat for a year tastes and extracts differently from a fresh one. Ask for the production or packing date, not only the best-before date.
Practical mechanisms that keep drift inside tolerance:
- Tolerance bands, not target points. Specify a range for extract yield, colour and moisture, and set the reject boundary explicitly.
- A retained reference sample on both sides, with an agreed review point.
- Pre-shipment sample approval for the first several shipments, and for the first shipment of every new crop year afterwards.
- A lot-linked certificate of analysis for every delivery, with methods named. Our botanical CoA guide covers how to read one, and the traceability guide covers what your own records need to capture.
- Lot reservation for a launch. If you are running a large first production, reserving a single homogenised lot avoids a mid-launch profile change that your own customers will notice before you do.
The document pack a formulator should ask for up front
Before any recipe work starts, request these. If they arrive without friction, the supplier can support a beverage programme; if they arrive slowly and incomplete, that is your answer.
| Document | What you use it for |
|---|---|
| Technical data sheet | Species, plant part, origin, cut, moisture, bulk density, storage conditions |
| Lot certificate of analysis | Actual values against your specification, with test methods |
| Microbiological report | Process design and kill-step validation |
| Pesticide and heavy metal reports | Regulatory compliance and importer duty of care |
| Allergen and cross-contact statement | Your own allergen labelling and line planning |
| Botanical identity confirmation | Species assurance, especially for traded material |
| Packaging and shelf-life statement | Warehouse planning and stock rotation |
| Sample with lot reference | Bench extraction before commitment |
The single highest-value item on that list is the last one. A bench extraction on a sample that carries a real lot number, run on your own water and your own time-temperature profile, tells you more than any specification document. Ask for enough sample to run the trial twice.
Working with a botanical supplier rather than a trader
The practical difference is whether the person answering your email can change anything. A trader can source and re-sell; a processor can adjust the cut, sieve out fines, blend to a target, run an additional test, and tell you what the field conditions were this year.
Arovela processes botanicals at our facility in Sındırgı, Balıkesir, in a region where many of the herbs concerned are grown, and holds stock in a warehouse in Solingen, Germany for European customers. We operate under ISO 22000, ISO 9001 and ISO 27001, and we can supply bulk material to a defined cut and fines specification for extraction, or — if part of your range is a bagged consumer product — run tea-bag production on a line filling roughly 1,000 bags per hour in single-chamber, string-and-tag and enveloped formats, from 5,000 bags upward. What we will not do is tell you that every crop year tastes the same. It does not, and a supplier who says otherwise is describing a blend they have not disclosed. The individual botanicals we handle are described in our Turkish herbal tea and botanicals overview.
FAQ
What cut size should I ask for if I do not know yet?
Ask for samples in two or three cuts of the same lot — for example a coarse cut, a medium tea-bag cut and a finer cut — and run your own bench extraction on each with your intended time and temperature. Choose on the results, then write the winning cut into the specification with a particle size distribution and a maximum fines fraction. Deciding by sample is faster and cheaper than deciding by argument.
Do I need steam-treated botanicals for a cold brew product?
It depends entirely on whether your process has a validated kill step downstream. If it does not, and the product is chilled rather than shelf-stable, treated material is usually the safer route. Be aware that treatment can reduce aroma intensity, particularly in delicate florals, so evaluate treated and untreated material side by side before deciding.
How much lot-to-lot variation is normal?
Some is unavoidable, because the input is a crop. What is reasonable is that variation stays inside agreed tolerance bands for the parameters you care about, that it is disclosed rather than discovered, and that the supplier can explain the cause — crop year, harvest window, region. Variation without explanation is the problem, not variation as such.
Can a supplier blend to keep my profile constant across crop years?
Often yes, within limits, by blending lots and by holding inventory across the crop transition. This requires planning and volume commitment, because the supplier has to carry stock to do it. It works best when you specify the target as a measurable range rather than as "the same as last time".
What can I say on the pack about a herbal ingredient?
Describe the ingredient, the origin, the plant part and the sensory character freely. Anything that states or implies a health effect falls under the EU nutrition and health claims regime and may only be used if authorised, and the position for botanicals is particularly restrictive. Settle the claim question with your regulatory adviser before artwork, because reprinting packaging is far more expensive than rewording a brief.
Starting a beverage programme
The most efficient first step is a specification conversation rather than a price conversation: species, plant part, cut, fines limit, microbiological requirement, annual volume and the process you will run. With those in hand, a supplier can tell you within a day whether they are the right partner — and if they are not, that is useful too.
Send us the botanical list and the process you are designing around, and request a quote. Samples for bench extraction go out with lot references so that what you test is what you can buy.