Where UK Online Software Spending Goes Abroad

Where UK software spending goes abroad is not necessarily where the software is developed, hosted or delivered. A company in Manchester can buy a cloud subscription through a merchant registered elsewhere, receive support from a UK team and use a service running on infrastructure in several countries. The Office for National Statistics (ONS) offers a way to examine one part of that relationship: aggregated and anonymised Visa card spending by UK cardholders with foreign merchants. Its research released on 6 February 2026 reports that computer software stores accounted for 5.7% of the analysed online spending abroad in 2024, while the United States received 40.7% of the international online spending within that specific merchant category. These are precisely defined measures, not a census of the UK's software industry.

The crucial distinction is between the recorded destination of a payment and the commercial or technical origin of the product. The same card payment can tell researchers something about merchant geography without identifying the specific programme purchased, its licence rights, the provider's data-processing arrangements or the country that produced its code. This matters to small UK IT companies, managed service providers and businesses choosing foreign subscriptions. Clear statistical boundaries help them interpret trade evidence while keeping practical supplier due diligence grounded in the documents that govern their own purchases.

Reading the 2024 computer software merchant evidence

The ONS report Consumer card spending, e-commerce and digital trade insights in the UK: 2019 to 2025 separates card purchases made online from face-to-face purchases and distinguishes domestic from foreign merchant locations. Figure 7 considers six leading merchant category codes within online spending abroad in 2024. These categories include large digital goods merchants, airlines, lodging, computer software stores, travel agencies and digital goods games. Computer software stores represented 5.7% of all online spending abroad in the relevant ONS category analysis; that 5.7% does not indicate the fraction of all UK spending directed to software or the industry's contribution to GDP.

Within computer software stores, the ONS states that the United States accounted for a larger spending share than any other individual destination country, at 40.7%. A country-specific percentage must not be mistaken for the whole North American region. Equally, the report's 75.4% European share refers to online spending across all six selected merchant categories, not to software stores alone. Applying that 75.4% to the software category would be a serious denominator error. An earlier version of this article contained additional continent percentages for the software-only category without a separately recovered chart-data file; rather than perpetuate uncertain figures, this rewrite retains the independently readable ONS claims and clearly identifies their measurement bases.

That caution is more than editorial tidiness. An IT procurement manager might be tempted to use a continental payment share to judge which region supplies the most trusted or best-quality software. The research cannot establish quality, security, uptime or supplier competence. It can only illuminate how payment spending is distributed within a specified card dataset. Procurement decisions require independent product and contract evidence.

Why the location on a card transaction is not a software supply-chain map

ONS merchant location information comes from records maintained for card processing, with merchant acquirers responsible for submitting the relevant information. For a shop with a fixed counter the location will ordinarily follow where the transaction happens. For a merchant without a fixed location, such as an online business, it can reflect where the transaction took place or its principal business address. These rules produce a useful statistical attribute, but they do not identify the physical location of the engineers who wrote an application, the headquarters of its parent group, the cloud region used for every customer, or the recipient of each tax payment.

Consider a hypothetical UK company subscribing to an overseas design platform. Its invoice may be issued by a regional sales entity, customer support may be available in London time and the platform's production systems may operate across several geographic locations. A card transaction can be assigned to the recorded merchant while leaving the rest of those arrangements unknown. The reverse is possible too: a domestically recorded merchant may resell technology built abroad. Payment location is one commercial fact, not a complete organisational chart.

The merchant category code introduces another limitation. It identifies a merchant's primary business activity under card-network classification, not the precise item purchased. A software-store payment might cover a licence renewal, a development tool or another offering. Certain digital services may appear under different codes entirely. Without product-level records, a researcher cannot claim that the 40.7% US share corresponds to a specific set of subscriptions, software brands or classes of buyers. Category labels should be reproduced accurately, without converting them into more detailed market research than the source provides.

The Visa data's practical strengths and hard limits

Card records provide an unusually direct view of how payments are made, including whether a payment card was present. In the ONS dataset, a transaction treated as online can include internet or app purchases and other card-not-present channels. The dataset is aggregated and anonymised before reaching the statistical office, so it is not a register of identifiable individual buyers. This makes it valuable for observing broad patterns, particularly among merchant categories where remote payment is common.

Yet card-not-present is not the complete international statistical definition of digitally ordered trade. The OECD, IMF, UNCTAD and WTO's Handbook on Measuring Digital Trade, Second Edition, published on 28 July 2023, distinguishes transactions ordered over computer networks from those delivered remotely over computer networks. A telephone card payment may be card-not-present without meeting the digitally ordered definition. An enterprise may also place a digital order and pay later by bank transfer. Consequently, Visa's payment-channel measure is useful evidence, not a one-to-one digital-trade census.

Coverage also matters. Visa does not represent every payment scheme, cash purchase, bank transfer or alternative payment method in the UK. The ONS explains that its data cover part of consumer spending and should not be taken as exhaustive. It further warns that Visa uses statistical disclosure controls, which can suppress granular country observations to prevent identification of particular people or businesses. When detailed figures are not visible, an analyst must not assume that a country has zero relevant spending. The six-category analysis is a selected segment of the dataset rather than a complete itemised global ledger.

These limitations do not invalidate the research. They tell readers how far its results can travel. It is fair to say the 2024 Visa analysis identifies the United States as the largest individual foreign country destination for spending classified under computer software stores. It is not fair to turn that sentence into a national claim about where all software sold to UK businesses is made.

What digital-trade statistics can add, and what they cannot replace

The ONS's Developing UK digital trade statistics: 2026 update, released on 26 June 2026, sets out why official compilers combine several imperfect sources. It separates digitally ordered transactions, digitally delivered services and the role of online intermediary platforms. Software may fall within digitally delivered services when supplied over a network, but the way it was paid for does not establish its method of delivery. An online purchase of a physical product is digitally ordered and physically delivered. A digital service can be delivered remotely even when the relevant customer agreement was concluded through a different process.

In its developing household-import methodology, the ONS uses Visa-derived proportions applied to broader Bank of England card-settlement values to estimate certain digitally ordered imports. That modelling step is different from the directly observed share of computer software stores in the earlier Visa analysis. It rests on additional coverage assumptions and does not allow a reader to calculate the total national value of software imports merely by multiplying a software-merchant share by a household-import estimate. Doing so would combine different populations and assumptions without proof of comparability.

For a software firm, separating the concepts is helpful when describing its own products. The sales contract identifies who supplies the service, while technical documentation can explain how users receive it and where their information is handled. The invoice identifies the billing arrangement. Trade statistics describe a wider economic pattern. These sources answer related, but not identical, questions. None should be silently substituted for another in a commercial case study.

Practical supplier checks for UK buyers

When evaluating an overseas software subscription, start with the legal name of the contracting supplier, payment currency and the service or licence being purchased. Examine whether the price includes the relevant taxes and whether it changes on renewal. If exchange conversion applies, record the terms and who bears the currency risk. The payment statement may show a different name from the brand on the website, which makes matching invoices to subscriptions especially important for bookkeeping and fraud prevention. Do not infer from a UK-looking website that every contract is with a British entity.

Next, examine access and continuity. Confirm who owns the organisation's account, how users are added and removed, what happens when a licence expires and whether critical data can be exported. A managed service provider supporting client software should know who has authority to approve new seats, change a billing method and obtain support during an outage. Separate customer support arrangements from marketing claims about global availability. Support hours, response commitments and escalation rights need to be verified in the applicable service documentation or contract.

Data protection is a distinct check. The location of a merchant for card-processing purposes does not reveal all processing locations or the contractual safeguards applicable to customer records. For services holding sensitive information, review the actual provider terms, the nature of the processed data, approved access and any required international transfer arrangements. A regional spending chart offers no shortcut to this assessment. Nor does a US or European payment designation establish a product's security quality or legal suitability on its own.

Finally, retain a practical procurement record: approved business purpose, actual service name, contract entity, purchasing owner, renewal date, current charge, support route and exit plan. Compare those facts with the invoice and service documentation. The best outcome of reading international payment statistics is often a better set of questions, not a false sense of certainty about the supplier behind every transaction.

Presenting and using the evidence without misleading readers

A responsible citation identifies the ONS publication date, the reference year 2024, the Visa-based population, the meaning of online, the foreign merchant-location attribute and the software-store merchant category. The 5.7% figure is the share of international online spending accounted for by that merchant code. The 40.7% figure refers to the USA's share within that code. If a report mentions Europe's 75.4%, it must say this aggregates six leading merchant codes rather than computer software stores alone. These percentages share a publication and broad payment source, but their denominators differ.

Researchers should also distinguish observation year from release date. A February 2026 report describing 2024 transactions is not evidence that the same regional distribution persisted in October 2026. Later data can change because buyers alter purchasing behaviour, merchants reconfigure their billing relationships or source methods are revised. The ONS explicitly cautions that changes in reported merchant location may reflect acquirer reporting rather than consumer decisions. A careful article leaves space for these possibilities rather than inventing a simple trend or presenting a dated chart as real-time market intelligence.

Software vendors, procurement teams and IT consultancies can use this material to understand one important limitation of market statistics: payment geography is not the same as product geography. The verified headline evidence identifies significant international online software-merchant activity, but it does not rank vendors, measure security or forecast a firm's revenue. The useful response is to preserve source definitions, perform supplier-specific checks and avoid treating all overseas digital expenditure as one comparable market. The result is a more defensible account of where spending appears to go, and of what remains unknown.

Source record: ONS consumer card spending report (6 February 2026), Figure 7 and methodology; OECD/IMF/UNCTAD/WTO digital-trade handbook (28 July 2023); ONS developing digital trade statistics (26 June 2026).