The Art Market Has a Data Problem. The U.K. Is Beginning to Address It.

A new British classification for art and antiques auction houses may appear to be a technical statistical change. In fact, it raises a much larger question: how can an increasingly sophisticated global art market operate without equally sophisticated data infrastructure?

By ArtAccess

Alex May

Memory Compression, 2025 screen-based work, moving image, audio-visual series

512 × 512 pixels Duration: approximately 5–10 minutes

The art market has spent decades becoming more global, more financialized, and more technologically sophisticated. Its data infrastructure has not always evolved at the same pace. That discrepancy is becoming increasingly difficult to ignore.

In September 2026, Artnet News reported on an important, if technical-sounding, development in the United Kingdom: the introduction of a dedicated Standard Industrial Classification (SIC) category for auction houses dealing in art, antiques, and antiquities.

The measure forms part of the U.K.'s revised system for classifying economic activity.

According to the U.K. Office for National Statistics (ONS), the Standard Industrial Classification framework provides a consistent structure for the collection, presentation, and analysis of economic data, allowing meaningful comparisons across industries and over time. Under the revised framework, UK SIC 47.921 specifically identifies the activities of retail auctioneering houses for artworks, antiques, and antiquities.

The classification itself will not standardize individual auction results. It will not determine how an auction house records a hammer price, how a guarantee is described, or how databases reconcile attribution, medium, provenance, or condition. Its immediate purpose is narrower: to make the economic activity of the auction sector more clearly identifiable within official statistics.

According to the ONS, SIC 2026 was finalized in April 2026 but will be introduced gradually, with its earliest anticipated use in the U.K. National Accounts in the 2031 Blue Book. The Art Newspaper, which reported on the initiative following the Art Business Conference in London, described it as an important step toward producing stronger statistical evidence of the auction sector's economic contribution.

Why should the art world care about an industrial classification code?

Because what an economy cannot classify properly, it struggles to measure properly. And what an industry cannot measure, it struggles to represent—to governments, financial institutions, researchers, investors, and ultimately to itself.

The development therefore exposes a much larger structural issue within the art market: one of the world's most valuable markets for unique assets still lacks a universally consistent language for much of the data it produces.

A $59.6 Billion Market Still Searching for a Common Language

The scale of the issue becomes clearer when considered against the size of the market itself.

According to the Art Basel and UBS Global Art Market Report 2026, authored by Dr. Clare McAndrew of Arts Economics, global sales of art and antiques reached an estimated $59.6 billion in 2025, rising 4% year-on-year after two consecutive years of declining values. Public auction sales increased 9% to approximately $20.7 billion, while dealer sales reached $34.8 billion. The United Kingdom remained the world's second-largest art market, accounting for 18% of global sales by value.

This is not an economically marginal industry.

Yet measuring it remains unusually complicated.

The market is distributed among auction houses, galleries, private dealers, advisors, online platforms, art fairs, artists, estates, and private transactions. Public auctions provide one of its richest sources of observable transaction data, while substantial parts of the dealer and private-sale markets remain confidential.

Even within the observable market, an auction result is not simply a price. It is a collection of information: artist attribution, title, date, medium, dimensions, provenance, exhibition history, literature, estimate, hammer price, buyer's premium, currency, sale location, guarantee status, condition, and whether the work actually sold.

Data becomes powerful only when its definitions are sufficiently consistent to allow meaningful comparison. That is why the U.K.'s new classification matters beyond Britain.

It illustrates a basic principle that the art industry has yet to apply comprehensively to its own transaction data: Standardization is the infrastructure that makes information comparable.

The ArtAccess View: Transparency Begins With Structure

The art market is often described as opaque. But opacity is only part of the problem.

A less discussed problem is fragmentation. There is now an enormous quantity of art-market information available. Auction houses publish catalogues online. Price databases contain decades of sales records. Market reports analyze millions of transactions. Digital platforms distribute information internationally within seconds. The challenge is increasingly not the absence of data, but the difficulty of making disparate data speak the same language.

Consider something as apparently straightforward as price. A published auction result may refer to the hammer price or to a price including the buyer's premium. Currency conversions may be calculated using different exchange rates. Works that fail to sell, are withdrawn before auction, or are sold after the auction may be treated differently across datasets.

The same fragmentation applies to objects themselves. Artist names can appear in different forms; attribution terminology varies; mediums and dimensions are not always recorded identically; and historical records are often less complete than contemporary ones. None of these differences makes the underlying information useless. But collectively, they introduce friction into comparison.

A genuinely standardized auction-data architecture would therefore go substantially further than the new British SIC classification. It could establish common definitions for the principal fields through which transactions are recorded and exchanged, while preserving the specialist information necessary to understand individual works.

For ArtAccess, the case for doing so is compelling.

Better Data Produces Better Price Discovery

Auction results are indispensable because they provide something comparatively scarce in the art market: observable transaction evidence. Historical results allow collectors, advisors, appraisers, lenders, insurers, and auction specialists to establish comparables. But the quality of a comparable depends on the quality of its underlying data.

A $2 million auction result is analytically different depending on whether $2 million represents the hammer price or the price including buyer's premium. A painting attributed directly to an Old Master is not equivalent to one catalogued as "workshop of" the same artist. Two works of similar dimensions by the same artist may occupy entirely different positions within that artist's market because of date, series, provenance, condition, exhibition history, rarity, or subject matter.

Standardization does not eliminate those distinctions. If properly designed, it makes them easier to identify. This matters because greater comparability can improve price discovery without pretending that artworks are interchangeable commodities. That distinction is essential.

The objective should never be to make art behave like equities. It should be to make the evidence surrounding an art transaction sufficiently intelligible that market participants understand exactly what they are comparing.

Better Information Can Lower the Market's Barrier to Entry

Information asymmetry has long been one of the art market's defining characteristics.

Experienced collectors and professionals accumulate knowledge through relationships, databases, auction catalogues, archives, private conversations, and years of market exposure. A new collector enters the same marketplace without that informational infrastructure. The consequence is not merely inconvenience. It can affect confidence and, ultimately, participation.

The question "Is this a fair price?" is considerably harder to answer in art than it is for most consumer goods or financial assets.

Standardized historical information cannot provide a definitive answer, because artworks are heterogeneous and valuation remains partly qualitative. But it can narrow the informational gap.

A collector who can clearly distinguish asking price from previous transaction prices, identify genuinely relevant comparables, see how frequently an artist's work appears at auction, and evaluate whether similar works have historically sold above or below estimate is in a stronger position to make an informed decision.

This is particularly important to ArtAccess's broader philosophy of accessibility.

Access to art should also mean access to the information necessary to participate intelligently in its market.

Transparency is therefore not simply about revealing more prices. It is about making existing information legible.

Standardized Data Could Strengthen the Art Market's Institutional Credibility

The British development demonstrates another consequence of poor classification: an industry can be economically significant while remaining statistically difficult to see.

The Office for National Statistics describes SIC precisely as a system for creating consistency across economic statistics. A dedicated classification for auctioneering in art, antiques, and antiquities should therefore make the sector's activity more visible within future economic measurement.

That matters for taxation, trade policy, cultural funding, employment, regulation, and government advocacy.

The same principle increasingly applies beyond government. The relationship between art and wealth management has deepened considerably. The Deloitte Private and ArtTactic Art & Finance Report 2025 found that 51% of surveyed wealth managers now offer art-related services, compared with approximately one-quarter when the report began tracking the field in 2011. Deloitte also estimates that approximately $992 billion in art and collectibles could change hands through wealth transfer over the coming decade.

As art increasingly intersects with estate planning, lending, insurance, wealth management, and family-office strategy, the demand for reliable information will only intensify.

Financial institutions do not require art to become a standardized asset. They do, however, require the information surrounding it to become increasingly rigorous.

Artificial Intelligence Makes Data Quality More Urgent, Not Less

There is another reason standardization matters now: artificial intelligence.

The art market is entering an era in which increasingly sophisticated systems can analyze enormous quantities of historical information—auction records, images, provenance, artist markets, exhibition histories, and pricing patterns—at speeds no human analyst could replicate.

But sophisticated models do not solve poor data architecture. They amplify whatever architecture they are given. If one database records premium-inclusive prices while another records hammer prices, if artist identities are inconsistently resolved, if currencies are improperly normalized, or if withdrawn and unsold lots are treated inconsistently, computational scale does not eliminate the problem.

It scales it.

The future value of AI to art-market analysis will therefore depend not only on better algorithms, but on better inputs: structured records, consistent definitions, transparent methodologies, and clear distinctions between observed facts and modeled estimates.

In this sense, data standardization is not an alternative to expertise. It is what allows technology to use expertise responsibly.

The Danger of Making Art Look More Precise Than It Is

There is, however, a compelling argument against taking standardization too far. The art market deals in fundamentally heterogeneous objects. A Basquiat is not simply another observation in a Basquiat dataset. Two paintings created months apart can differ radically in quality, historical importance, provenance, imagery, condition, and desirability. A major museum exhibition can alter the critical understanding of an artist. Distinguished provenance can materially affect the perception of a work. A single determined bidder can transform an auction result.

Moreover, auction prices do not emerge in a vacuum. Guarantees, irrevocable bids, reserves, withdrawals, and other mechanisms can influence the conditions under which works reach the block and sell. The importance of these structures was particularly visible during the May 2026 New York auctions. Christie’s highly successful S.I. Newhouse sale, for example, was 100% sold, but The Art Newspaper reported that the entire sale was backed by third-party guarantees. This is precisely why a recorded sale price cannot always be understood in isolation from the structure of the transaction.

A standardized database can record that a guarantee existed. It cannot make the economic significance of that guarantee disappear, nor can auction data describe the entire art market.

The Art Basel and UBS Global Art Market Report 2026 estimates dealer sales at $34.8 billion in 2025, compared with $20.7 billion in public auction sales. Auction houses reported a further sum of just under $4.2 billion in private sales. Significant additional activity takes place privately through galleries, dealers, advisors, and collectors, where transaction prices are rarely made public. The most immaculate dataset of public auction results would therefore still describe only part of the market.

This creates perhaps the greatest danger of data-driven analysis: false precision.

When information is cleanly structured and presented through increasingly sophisticated models, estimates can begin to look like facts. A valuation of $1.8 million appears more authoritative than a range of $1.5 million to $2 million, even when the underlying evidence cannot justify that degree of certainty. Standardization should therefore increase the quality of judgment—not replace judgment itself.

What Should Be Standardized—and What Should Not

The distinction is ultimately straightforward. The art market should standardize facts, not value judgments.

Transaction date, auction house, location, currency, estimate, hammer price, premium-inclusive price, lot status, dimensions, medium, edition information, guarantee status, and clearly defined transaction characteristics can be structured.

Historical significance, aesthetic quality, rarity within an artist's practice, institutional importance, and the cultural meaning of an artwork cannot be reduced to the same taxonomy without losing something essential.

This is where connoisseurship and data should meet rather than compete. The most mature version of an information-driven art market is not one in which algorithms determine what an artwork is worth. It is one in which reliable data handles what can be measured, allowing specialists to concentrate more intelligently on what cannot.

From More Data to Better Data

The new British classification will not transform auction transparency overnight. Nor is that its purpose. UK SIC 47.921 is fundamentally an instrument of economic classification and measurement, rather than a universal standard for recording individual auction transactions. As Artnet News and The Art Newspaper reported in September, its importance lies in making the auction sector's contribution to the British economy more visible and measurable.

Its broader significance lies in the principle behind it. Before an industry can be properly analyzed, its activity must first be properly identified. Before data can become evidence, it needs structure. That principle deserves to travel further.

The art market has already experienced the first stage of its information revolution. Auction results that once circulated largely among specialists are now searchable around the world. Catalogues are digital. Market reports are increasingly sophisticated. Vast historical datasets can be interrogated within seconds.

The next stage should not simply be about producing more data. It should be about producing better data: interoperable where appropriate, transparent in methodology, precise in definition, and honest about its limitations.

For collectors, that could mean greater confidence. For advisors and appraisers, stronger comparative analysis. For policymakers, a clearer understanding of the industry's economic importance. For financial institutions, more reliable evidence. And for the technology companies increasingly entering the art ecosystem, a stronger foundation on which to build analytical tools.

The paradox is that better standardization need not make the art market less human.

It may do the opposite. By creating greater certainty around the things we can measure, we become more intellectually honest about the things we cannot: quality, significance, taste, cultural importance, and ultimately the human judgment that determines why one work matters more than another.

The future of the art market does not require turning art into data. It requires making the data surrounding art worthy of the market it is meant to describe.