The Service Economy: Understanding Trade Using AI and Web Data.
Trade — a cornerstone of the global economy
Whilst the global economy remains volatile and subject to macro-level factors such as war, the climate emergency, cutting-edge technology disruption and political instability, the importance of trade remains a constant. The effect of trade is multifaceted — it shapes dimensions of economic and political relationships and binds countries together that would otherwise have minimal association. As such, trade is often a central pillar of economic policy, reflecting the dividends at stake. Governments therefore seek to secure opportunities for the global flow of domestic products, services, ideas and investment, as a key symbol of economic competitiveness and geopolitical influence.
In the UK, trade maintains a very high profile in Government and across the business and investment community. This is evident in an eagerness to promote the significance of key sectors, and economic strength and showcase global trading ambitions, which have become more pronounced since the UK left the European Union. The Government is pushing hard to promote UK PLC, creating a bolder growth narrative expected to underpin opportunities for British exporters of goods and services and a new wave of foreign companies investing in the UK.
As with other policy areas, understanding the nature, scale and flow of trade, relies on good information. Currently, the focus is on gathering intelligence for official datasets, leveraging primary research methods and administering large-scale surveys. This has historically emphasised product and goods data as a tangible measure of trade, for example, there are approximately 280k distinct companies in the HMRC trade data (the publicly available list of UK importers/exporters). But in light of economic shifts, there has been an emergence of services trade data. This is especially important for the UK, as it has become a services-orientated economy, with powerhouse sectors propelling this transition. At a high level, recent data highlights the UK’s service economy prowess in stark terms — it is the second largest services exporter in the world, with the UK services industries accounting for 81% of domestic GVA over the period October-December 2023 (DBT, Business and Trade Facts and Figures, April 2024 / UK Parliament, House of Commons Library, Service industries: Key Economic Indicators, May 2024).
Services trade data: advancements in reporting but still limited compared to goods trade data
The UK is blessed with a wide array of open data, mostly published by the Office for National Statistics (ONS) and HMRC. These provide a wealth of information, with many datasets produced at regular intervals (i.e. annual), enabling longitudinal analysis and the tracking of change. A prime example of this, with a trade focus, is the Annual International Trade in Services publication, which is the primary basis for Governmental reporting on services trade. The data is emblematic of official sources — collected via large-scale surveys, generating results that can be categorised by high-level sector and cross-referenced with other datasets. The data is prepared on the change of ownership principle, and so reflects a sub-set of the economy, helping to assign a value to transactions. This introduces some inherent limitations associated with the research approach:
The data is based on sample responses, thereby reflecting a proportion of companies within the economy.
Some service sectors are omitted, including Travel and Transport, Banking and other Financial Institutions and Higher Education.
Despite this, the data paints an interesting picture of services trade and makes an important contribution to national accounting processes (including being a condition of the UK’s International Monetary Fund membership). The latest data release illustrates the value of this intelligence by showing the global standing of UK services trade:
Beyond this, the data exposes additional layers of insight within specific sectors, including those that are of national importance and service-orientated. An example of this is the Digital Sector — a longstanding focus on growth and representation of the UK’s cutting-edge economy. Here, a deeper analysis has been conducted by the Department of Culture Media and Sport (DCMS) using common data, but at a sector level (adopting a combination of SIC codes). Key messages from this analysis show that in 2021 (DCMS, Digital sector economic estimates: Trade, 2021 — technical and quality assurance report, Updated May 2024):
There were £59 billion in digital services exports (19% of all UK service exports).
Digital services exports have been consistently higher than imports of services between 2016 and 2021 — denoting a trade surplus.
The USA (£20.9 billion) was the largest market for digital services exports, followed by Ireland (£4.3 billion).
The data helps to reinforce messages around economic recovery, the UK’s status as an economic powerhouse in digital services and how integral and cutting-edge these sectors are to prosperity. However, the data also exhibits constraints, which can impede policymakers and obscure deeper layers of intelligence. Key examples of this include the ability to identify specific businesses that are exporting services, the nature of the services and the destination of these exports.
Uncovering trade intelligence with AI and Web data
Using our AI capability that deep reads the web to research companies, we have gathered trade data indicators across sectors and geographies. Moreover, the capability has been particularly helpful in uncovering evidence of trade in services-orientated sectors (where there a significant gaps in evidence from official sources).
On a practical level, there are recent examples of us investigating trade for clients across the UK, including within sectors that have a services focus. This includes various projects for the Department for Business and Trade, the North East Local Enterprise Partnership and the Great South West Pan-Regional Partnership, to name a few. The emphasis of these projects has been that of a common goal — to identify companies within sectors of interest that are trading abroad, inclusive of service and product exports.
A novel but tested approach
Using the glass.ai sector taxonomy, which provides a contemporary classification of the UK and global economy, it is possible to look at industries that are services-orientated, in greater detail.
Once we have identified the services companies, the next step involves deep crawling the websites of the selected businesses to uncover indicators or signals of trade with specific countries, including:
Does the company have offices abroad? if so, what are the locations?
Does the company have people/employees abroad?
Does the company have clients, partners, suppliers, or JVs abroad?
Does the company mention specific countries or geographies in press releases or external news?
Is the company listed in some of the official sources (e.g. HMRC list of importers/exporters)?
These are just some examples of indicators, signals and evidence of trade our AI will aim to uncover. It will also aim to discover any trade associations and/or sector sources that might help us identify good candidates for the research. This novel strategy where we apply our AI to deep-read UK business websites, social media and other sources has worked very well in many projects that required building databases of UK services companies trading with other countries. The objective is not to replace official data, but rather augment and complement it with detailed evidence sourced from the open web.
Bringing this into context by considering the Digital Sector, we recently looked at lower-level segments that are most likely to be services orientated and then we gathered trading evidence in greater detail. Below we illustrate trade insights about the UK’s digital sector that wouldn’t be captured through official data collection methods:
One of the most insightful facets of using a deep crawling approach is obtaining company-level intelligence, as a foundation for deeper exploration. Here we can observe the total number of UK companies within the IT & Technology and Services sector that are trading services internationally, expressed across several ‘signals’. In the infographic are some examples — interesting by virtue of their UK location (and global footprint), size, business focus, service offerings and the strength of trade signals observed (3 or more signals detected by the crawl). This data can be traced at a granular level too, providing direct links to provenance and contextual relevance, but can also aggregated to tell a composite picture of company trading, or indeed, at a larger denomination, at a sector level.
For researchers and policymakers, this approach offers tangible advantages, such as being able to identify trade-intensive companies, consider the spatial dimension of global trading relationships, and gain a fuller view of services trade, across a wider range of sectors, including those not identifiable via traditional methods (i.e. survey-based using SIC codes). It also offers a degree of flexibility not possible within static databases, where signals can be discovered based on specific areas of research interest. Moreover, the use cases for this data are increasingly relevant and reflect the delivery of important work for public sector clients:
Making and reinforcing the case that emerging and new sectors are helping to drive trade activity and economic growth.
Identifying countries that could and should be the target for sector-specific trade policy, promotion and intervention.
Pinpointing companies within sectors that are strong candidates for support and expansion into international markets.
Developing sample frame datasets that allow direct engagement with target companies and sectors to understand their growth plans.
The value of bespoke trade and sector intelligence and company-level insight becomes even more apparent when considering the global expansiveness of open web data. Targeting the crawl towards foreign countries or target markets, equivalent, rich data can be captured. This has several applications for end users — such as considering the existing influence of companies within the UK, or their potential role as a driver of growth and inward investment. By creating scalable datasets, answering granular trade questions becomes not only possible but a reality.
AI and the Web becoming integral to trade research
Trade remains a high priority for Governments and jurisdictions across the world, and an AI and data revolution is helping to understand the complexities of these relationships, at a more granular level. Whilst the development of official trade datasets represents a positive step forward, especially in the context of understanding services trade, this approach only provides limited evidence. The opportunity for AI and web crawling is already here to augment official sources with new layers of evidence.
Given our key role in several trade research projects, at glass.ai we are well-placed to help public sector bodies with new assignments that require a granular level of trade data, especially across service-intensive sectors not captured in official data. We can also support investigation into bespoke sectors, or emerging areas of the economy, using a consistent approach that provides data/evidence for policy development, investment and place promotion. We look forward to collaborating with public sector clients, consultancies and market research firms on assignments that leverage the full value of this novel but tested approach.