What a honey authenticity database is

Modern authenticity testing does not read a verdict off an instrument. It compares a sample against a collection of honeys believed to be genuine — and the quality of that collection decides what the result is worth.

Quality, testing and standardsReviewed 2026-08-28

What the instrument actually tells you

Authenticity analysis divides into two kinds. Targeted analysis looks for something specific — a named marker compound, a particular syrup residue — and knows in advance what it is asking. Untargeted analysis records a broad signal instead: a nuclear magnetic resonance spectrum, an infrared profile, a mass-spectrometric fingerprint. Nothing in that signal is labelled genuine or adulterated. It becomes an authenticity finding only when it is compared against a body of data from honeys believed to be real.

This is the step most reporting skips. A headline saying a laboratory found a honey to be adulterated is usually describing a sample whose profile sat outside the range a database holds for honeys of that description. That is a real and useful finding. It is not the same kind of statement as a measurement of added syrup, and it inherits every weakness of the collection it was compared against.

Why the transparency problem became a legal problem

The largest reference collections were built commercially, at considerable cost, and are not published. That is a defensible commercial position and it creates a specific difficulty: if the comparison set cannot be examined, neither the representativeness of the comparison nor the interpretation drawn from it can be independently checked. Findings resting on those databases have supported significant commercial and enforcement decisions, and their opacity has been identified as a barrier to their acceptance and a cause of legal dispute.

Defra, the Food Standards Agency, Food Standards Scotland and the Government Chemist have been working on the scientific side of this since 2018, following a Government Chemist seminar on honey authenticity in 2019. The route they took was not to build a public database or to reject proprietary ones. It was to write down how such a database should be interrogated when its conclusions are challenged, with safeguards for confidential data so that scrutiny does not require publication.

What a database is interrogated for

The framework published in November 2025 groups the questions under five headings. They are worth knowing even outside a dispute, because they are also the reasons a database-based result can be right, wrong, or simply not applicable to the sample in front of it.

ScopeWhat was this database built to answer, and was that decided before it was assembled or discovered afterwards? A collection begun to detect one kind of adulteration and later used for geographical origin is being asked a question it was not designed for, and stating what a database is not suitable for is treated as part of declaring its scope.
CompositionWhat is actually in it, and on what basis were samples admitted or excluded? A reference database is only as authentic as its criteria for deciding that a sample was genuine in the first place — a circularity that has to be broken by evidence outside the analysis, such as collection directly from a known beekeeper.
MetadataWhat is recorded alongside each sample: botanical source, geographical origin, bee species, year and season, processing, and the traceability of the sample itself. Without this the collection cannot be subset, and a sample can only be compared against the whole thing rather than against the honeys it actually claims to be.
RepresentativityDoes the collection capture the natural range of what it purports to cover, including seasons, years, regions and production practices? If it claims to represent honey globally, does it reflect where honey is actually produced? Is it curated and reviewed as the world changes?
ValidationWere the analytical methods properly validated before being used, and are they accredited to an international standard such as ISO/IEC 17025? A database is only as good as the measurements that populated it.

What a database match can and cannot show

Reading a database-based authenticity result
Can help show	Cannot by itself establish
That a sample's profile is unusual relative to honeys believed genuine and comparably described	That the sample is adulterated. Rarity is not fraud, and an under-represented honey is rare in the database for innocent reasons
That a specific known adulterant signature is present, where the method is targeted at one	That no adulterant is present. Absence of a match is absence of the signatures held, not absence of syrup
That a sample is inconsistent with the region or flower it is declared as, where those honeys are well represented	Where the honey did come from. Excluding a claimed origin is not the same as establishing the true one
A basis for further, targeted investigation and for asking the supply chain questions	Intent, or who in a chain of a dozen handlers was responsible

Where this is going

The framework is a method rather than a rule: it carries a disclaimer that it represents its authors' views and not government policy, and its own conclusion is that database scrutiny needs standardised approaches while avoiding a counsel of perfection no real database could meet. Its authors expect it to be generalised to authenticity databases for other foods.

Two things are worth watching. The first is whether a mechanism for confidential data transfer is established, which is what would let a database be examined without being published. The second is the European side: Directive (EU) 2024/1438 empowers the Commission to introduce harmonised methods for detecting sugar adulteration and a uniform methodology for tracing honey origin, and if those arrive they change what a reference collection has to be able to do.

The claims on this page

Every substantive claim is placed on a tier, and the top two tiers must also state what they are not claiming.

Established evidence

Supported by systematic reviews, meta-analyses or clinical/regulatory guidance. The claim would survive a careful reader checking it.

Established evidence

The proprietary databases used to interpret non-targeted honey authenticity tests are largely unpublished, and a UK framework now exists for interrogating them rather than accepting their results on trust.

A project commissioned under Defra's Food Authenticity Programme and jointly funded with the Government Chemist published a framework on 28 November 2025, developed by an expert working group that met between October 2023 and December 2024. Its starting point is that the use of proprietary databases to interpret results, particularly from nuclear magnetic resonance, has been identified as a significant barrier to acceptance, because opacity makes representativeness and interpretation difficult to check, and that this has led to legal disputes. The framework sets out how to appraise a database's scope, composition, metadata, representativity and method validation, with safeguards for the owner's confidential data.

What this does not claim: This is not a claim that any named database is inadequate, that results derived from these databases are wrong, or that the laboratories using them act improperly. The framework's own position is that a database which is fit for purpose gives a reliable guide to authenticity; the point is that fitness for purpose is a question that can be asked and answered rather than assumed. It is also not government policy, and it does not create a legal requirement on anyone.

Framework for interrogation of honey authenticity databases (IHAD) — Government Chemist and Department for Environment, Food and Rural Affairs (2025) · Regulatory guidance Link

Sources

Framework for interrogation of honey authenticity databases (IHAD) — Government Chemist and Department for Environment, Food and Rural Affairs (2025) · Regulatory guidance Link
Understanding honey fraud and the role of authenticity testing — Government Chemist (2026) · Regulatory guidance Link
Directive (EU) 2024/1438 amending the 'breakfast directives', including honey — European Parliament and Council of the European Union (2024) · Regulatory guidance Link
Analytical approaches to honey authentication — Peer-reviewed food-analysis literature · Peer-reviewed article

How sources are selected and weighted is set out in the sources and evidence policy.

In short

A reference database defines the natural variation of genuine honey; a result is a comparison against it, not a direct measurement of fraud.
The large honey authenticity databases are commercially held and largely unpublished, which has made their conclusions contestable in law.
A database can only speak about honeys it adequately represents, so a genuine honey from an under-sampled region can look anomalous.
Metadata — where, when, from what and from which bee each reference sample came — is what makes a database interrogable at all.
The Government Chemist and Defra published a framework in November 2025 for scrutinising these databases rather than taking them on trust.

This guide answers

What is a honey authenticity database, and why does a database match not settle the question?

Tools that use this

Honey finder
Which honey should I buy for what I actually want to do with it?
Flavour selector
I like this taste — which honeys have it?
Honey comparison
How do these honeys actually differ, beyond colour?

What a honey authenticity database is

What the instrument actually tells you

Why the transparency problem became a legal problem

What a database is interrogated for

What a database match can and cannot show

Where this is going

The claims on this page

Established evidence

The proprietary databases used to interpret non-targeted honey authenticity tests are largely unpublished, and a UK framework now exists for interrogating them rather than accepting their results on trust.

Sources

Read next

Tools that use this