"Concentrated natural extract" is a description, not a specification. It tells you that plant material went in and something denser came out. It does not tell you which species, which plant part, how much raw material stands behind a kilogram of powder, what solvent touched it, how much of that kilogram is carrier rather than extract, or how much the next lot is allowed to differ from this one. Every one of those unknowns is a place where two quotes that look comparable are describing genuinely different products.
This guide is written for buyers who have to turn a vague brief into a document a supplier can be held to — formulators at supplement and cosmetic brands, technical buyers at ingredient distributors, and QA staff who inherit whatever procurement agreed. It walks through the fields a concentrated-extract specification needs, explains what each one prevents, and shows how to write a batch-to-batch tolerance that means something. If you want the deeper theory behind the concentration figure itself, read it alongside the extract standardisation and ratio guide, which unpacks the difference between a ratio and a marker in detail.
Why a loose specification is expensive
An underspecified extract does not fail loudly. It fails quietly, three lots in, when a formulation stops behaving, a customer audit asks a question nobody can answer, or a lab result arrives that the contract never anticipated.
The three characteristic failures are worth naming, because each maps to a specific missing field:
Substitution you cannot prove. Without a botanical identity field — Latin binomial, plant part, and an identity test — you have no contractual basis to reject material that is the right family but the wrong species, or the right species but the wrong part. Leaf standing in for root, or aerial parts standing in for flower, is not necessarily fraud; sometimes it is a supplier honestly interpreting a vague order.
Dilution you cannot detect from the price. A powder that is largely maltodextrin will always undercut a powder that is largely native extract. If the specification does not state the percentage of native extract, the cheaper offer is not cheaper — it is a different product, and the comparison was never valid.
Variability you cannot plan around. Plants vary by season, region and harvest year. That is normal and unavoidable. What is avoidable is discovering the size of the variation after you have designed a formulation around one exceptional lot. A tolerance band written into the specification converts an unpleasant surprise into a known operating range.
The anatomy of a concentrated-extract specification
A workable specification is not long. It is complete. The table below lists the fields that carry weight, what each one commits the supplier to, and the failure it prevents.
| Field | What it must state | What it prevents |
|---|---|---|
| Botanical identity | Latin binomial, plant part, country/region of origin | Species and plant-part substitution |
| Concentration basis | Native drug-extract ratio, stated as native | A ratio quietly measured after carrier |
| Standardisation | Named marker, percentage as a minimum or range, assay method | An uninterpretable "5%" with no method |
| Extraction process | Solvent(s) used, process type, whether decolourised or deodorised | Undeclared solvents and hidden refining |
| Residual solvent | Named solvents with maximum limits and test method | Solvent carry-over discovered at customer audit |
| Carrier / excipient | Identity and percentage of every added substance | Apparent concentration inflated by bulking |
| Moisture / loss on drying | Maximum, with method and temperature | Water sold as extract; caking; microbial growth |
| Ash / acid-insoluble ash | Maximum | Mineral or soil contamination |
| Particle size and form | Mesh or micron distribution, powder/liquid/oleoresin | Blending and dosing problems in production |
| Solubility and appearance | Behaviour in the intended medium, colour and odour range | Formulation failures nobody specified against |
| Heavy metals | Named elements, maximum each, method | Lots rejected at import |
| Microbiology | Named organisms and counts, method | Non-conforming finished product |
| Pesticides / contaminants | Screening scope and reference limits | Regulatory exposure carried by the buyer |
| Shelf life and storage | Months from manufacture, temperature and humidity conditions | Disputes over degraded material |
| Packaging | Primary container, liner, net weight, labelling of lot and date | Damaged or untraceable receipts |
| Batch-to-batch tolerance | Permitted variation lot to lot and the out-of-spec procedure | Every lot being a fresh negotiation |
Read that table as a checklist rather than a template. Not every line carries the same weight for every material — particle size matters enormously for a dry powder going into a tablet blend and hardly at all for a liquid oleoresin — but every line should be either filled in or consciously marked as not applicable.
Ratio versus standardised-to-marker: choose deliberately
The single most consequential choice in the specification is how you define concentration, and buyers routinely treat two different systems as if they were grades of the same one.
A ratio specification — written as 4:1, 10:1 and so on — describes how much dried raw material was used to make one unit of extract. It is a statement about process, not about content. Its virtue is that it is honest about what a whole-plant extract is: a concentrated version of the entire botanical, with its natural profile broadly intact. Its weakness is that it says nothing about the level of any particular constituent, so two 10:1 extracts of the same species from different harvests can differ substantially.
Two things must be pinned down for a ratio to be meaningful. First, whether it is native — the ratio of starting material to extract before any carrier is added — or adjusted after bulking. Second, the specification must state that the ratio is native in words, because the number alone does not distinguish the two.
A marker specification — "standardised to X% of named constituent" — fixes the extract to an assayed content of a specific compound. Its virtue is reproducibility: batches are released to a target, so your formulation sees a consistent input. Its weakness is that the target can be hit in ways you did not intend, including by adding an isolated compound to weak extract. A marker figure is only as trustworthy as the assay method named beside it and the identity testing behind it.
| Approach | Best used when | Must also specify | Weak point |
|---|---|---|---|
| Native ratio (DER) | Whole-profile extracts; no single dominant constituent | That the ratio is native; carrier percentage; a fingerprint | Constituent levels can drift lot to lot |
| Standardised to marker | Formulation needs a consistent assayed input | Marker name, minimum or range, assay method, reference standard | Target can be met by fortification |
| Both together | Most technical purchases | Native ratio, marker, method, carrier, fingerprint | Longer specification, more testing cost |
For most technical purchases the right answer is both: a native ratio that describes the material, a marker with a named method that describes the release gate, and a carrier declaration that reconciles the two. The COA reading guide for botanical ingredients covers how those three numbers should appear on a certificate and what an inconsistency between them indicates.
Solvent, process and residual solvent
An extract is defined as much by how it was made as by what it contains. Specify the process, not just the outcome.
State the extraction solvent or solvents by name — water, ethanol, an ethanol-water ratio, supercritical carbon dioxide, glycerine, a vegetable oil. This is not a courtesy; solvent choice changes which constituents are pulled out of the plant, which is why an ethanol extract and a water extract of the same botanical are chemically different materials that happen to share a name. The comparison between the two dominant industrial routes is set out in the CO2 versus ethanol extraction guide.
State also whether the extract has been decolourised, deodorised, filtered or de-bittered, because each of these steps removes something and none of them appear in a ratio or a marker figure.
Then specify residual solvent limits by name, each with a maximum and a test method. Extraction solvents used in food production are regulated in the EU under a dedicated directive, and the constituent-level and labelling framework a finished product sits inside is set by Regulation (EU) No 1169/2011 on food information to consumers. Whether your supplier tests for residual solvent as standard or only on request is a question to settle before the first order, not after the first audit.
A specification that names carbon dioxide as the solvent has a short residual-solvent section; one that names ethanol or a hydrocarbon needs a real one.
Carrier, moisture, ash and the physical fields
These are the unglamorous fields that decide whether the material works on your line.
Carrier and excipient declaration. Require the identity and percentage of everything added: maltodextrin, gum arabic, starch, silicon dioxide, tricalcium phosphate, any absorbent used to fix a liquid onto a powder. The specification should express the material as a percentage of native extract plus a declared remainder. Without this, an extract-equivalent calculation is impossible and price comparison is guesswork.
Moisture or loss on drying. State a maximum, and state the method and temperature, because different drying protocols give different numbers on the same sample. Moisture is not a cosmetic parameter: it drives caking, it drives microbial growth, and it is the most direct way to sell water at extract prices.
Ash and acid-insoluble ash. Total ash reflects the mineral content of the material; acid-insoluble ash is the classic indicator of soil and sand. Both belong on a specification for a plant-derived powder.
Particle size, bulk density and form. For powders, state a mesh or micron distribution and, if it matters, bulk and tapped density. For liquids, state viscosity and whether the material is a fluid extract, a soft extract or an oleoresin. Formulators discover these fields the hard way when a blend segregates or a capsule filler under-doses.
Appearance, odour and solubility. Write down the acceptable colour range and the behaviour expected in the intended medium — cold water, hot water, ethanol, oil. Colour drift between lots is normal for plant material and is only a dispute when nobody wrote down the range.
Contaminants: metals, microbiology, pesticides
Concentration concentrates everything, including what you did not want. This is the reason a contaminant section on an extract specification cannot simply be copied from the raw botanical.
Heavy metals. Name the elements to be tested rather than writing "heavy metals: complies". Lead, cadmium, mercury and arsenic are the conventional set; state a maximum for each, state the analytical method, and state that results are reported as figures rather than as the word "conforms". The mechanics of interpreting these results across botanical and dried-fruit categories are covered in the heavy metals guide.
Microbiology. Specify total viable count, yeasts and moulds, and the pathogen absences relevant to your finished product and its process. A dry powder that will not be heat-treated downstream needs a tighter microbiological specification than one entering a cooked application, and the specification should say which situation applies. The botanical microbial limits guide sets out how those limits are structured.
Pesticides and other contaminants. State the screening scope — a named multi-residue panel rather than an open-ended promise — and the reference limits being applied. Where the botanical carries a known category-specific risk, name it explicitly rather than relying on a general screen to catch it.
One further boundary belongs in the document: a specification describes composition and safety, not effects. Statements about what an ingredient does to health sit under Regulation (EC) No 1924/2006 on nutrition and health claims and are the finished-product brand's regulatory responsibility. Keep them out of the ingredient specification entirely; a supplier who volunteers them is telling you something about their documentation discipline.
Writing a batch-to-batch tolerance that means something
This is the field most often left blank, and the one that determines whether the specification survives contact with real agriculture.
Plant material is variable by nature. A specification that ignores this either gets quietly breached every season or forces rejection of perfectly good material. The fix is to state, for each measured parameter, three things:
- The limit — a minimum, a maximum or a range, with a unit and a method. Never "typically". A typical value is marketing; a limit is a commitment.
- The permitted lot-to-lot variation — for parameters where you care about consistency rather than only about a threshold, such as marker content, colour or particle distribution. A marker specified as a minimum allows an unlimited upside, which sounds generous until a lot arrives at double the usual strength and your dosing is wrong.
- The out-of-specification procedure — who is notified, within what window, whether retesting is permitted and on what sample, who pays for the retest, and what the disposition options are. Concession acceptance at an agreed adjustment should be an explicit option, because for agricultural material it is often the sensible commercial outcome.
Add two supporting mechanisms. Retention samples: require the supplier to hold a retained sample of every lot for a defined period, and hold your own. Without paired retains, an eventual disagreement about a result is unresolvable. Trend data: ask for the last several lots' results on the key parameters before you commit. A supplier who can show a stable trend has real process control; one who can only show the current lot's certificate is showing you a snapshot.
Finally, decide where the specification is verified. Supplier certificate only, supplier certificate plus periodic third-party verification, or full incoming testing on every lot — each is defensible, but the choice should be written down and reviewed as the relationship matures rather than drifting by habit. Running a properly structured sample round first makes that decision much better informed; the B2B sample order guide covers how to structure one so the sample actually predicts production.
From specification to quote
A specification is also a procurement tool. Suppliers who quote against a complete specification are quoting the same product as each other, which is the only condition under which comparing offers is meaningful. Suppliers who quote against "10:1 extract, best price" are each quoting a different product and the cheapest is usually the least like what you wanted.
Send, at minimum: species and plant part; the concentration basis you require, stated as native ratio and/or marker with method; solvent and any prohibited solvents; the maximum carrier percentage you will accept; moisture, ash, metals and microbiological limits; packaging and net weight; annual volume and call-off pattern; and the destination market, because destination determines which contaminant and labelling frameworks apply.
Arovela works this way by preference rather than as a formality: quotes are issued against a written specification, not against a product name, and the accompanying documentation is produced within management systems certified to ISO 22000, ISO 9001 and ISO 27001. Production sits in Sındırgı, Balıkesir, with a warehouse in Solingen serving EU customers, and the served markets are the EU and Ukraine. Where a parameter cannot be met, the answer is that it cannot be met — which is more useful than a quote that quietly reinterprets your specification downward. For adjacent sourcing context on the EU side, the botanical extracts sourcing guide for supplement brands covers supplier qualification and documentation flow.
FAQ
What is the difference between a ratio extract and a standardised extract?
A ratio extract states how much raw plant material produced one unit of extract, for example 10:1. It describes concentration achieved by the process but says nothing about the level of any specific constituent. A standardised extract is released against an assayed percentage of a named marker compound, which gives batch-to-batch consistency for that constituent but can, in principle, be achieved by adding isolated material to a weak extract. Most technical specifications should state both, together with the assay method and the carrier percentage, so the two figures can be reconciled against each other.
Why does the extraction solvent need to be on the specification?
Because solvent choice changes what is actually in the extract. Water, ethanol, ethanol-water mixtures and supercritical carbon dioxide each pull different constituent classes out of the same plant, so extracts sharing a botanical name can be chemically different materials. Naming the solvent also determines what residual-solvent testing is required: a carbon dioxide extract needs almost none, while an ethanol or hydrocarbon extract needs named limits and a stated test method. Specify any post-processing too, such as decolourising or deodorising, since those steps remove material without changing the ratio.
How much carrier is acceptable in a concentrated extract?
There is no universal figure, and that is precisely why the specification must state the percentage of native extract and the identity and percentage of every added substance rather than a generic allowance. Carriers such as maltodextrin, gum arabic, starch and silicon dioxide are legitimate and often necessary to make a hygroscopic extract flowable and dosable. The problem is undeclared carrier, which makes one offer look cheaper than another when in fact it contains less extract. Decide the maximum you will accept for your application and write it into the document.
What should a batch-to-batch tolerance clause contain?
Three elements: the limit itself as a minimum, maximum or range with a unit and a method; the permitted variation from lot to lot for parameters where consistency matters, such as marker content or colour; and the procedure when a lot falls outside specification, covering notification timing, retesting rights, who pays, and whether concession acceptance at an adjusted basis is available. Support it with retained samples held by both parties and a request for trend data across several recent lots, which reveals real process control far better than a single certificate does.
Can a supplier's certificate replace my own testing?
It can, if you have decided that it does and written that decision down. The three defensible positions are supplier certificate only, supplier certificate with periodic independent verification, and full incoming testing on every lot. Which is appropriate depends on the risk profile of the material, the maturity of the relationship, and what your own customers or auditors expect of you. What is not defensible is having no stated position, because that is how a certificate that says "conforms" instead of reporting a numeric result quietly becomes your quality system.
Send your species, plant part, concentration basis, solvent constraints, carrier ceiling, contaminant limits, packaging and destination market, and request a quote — we will quote against that specification, tell you plainly which parameters we can hold and which we cannot, and confirm the documentation that will accompany each lot.