Last year I had the pleasure of attending a meeting hosted by the Royal Academy of Engineers (RAEng), discussing the challenges of valuing the UK’s data economy. The meeting followed on from the Academy’s publication of Connecting data – Driving productivity and innovation, jointly created with the Institute of Engineering and Technology, and was attended by key thinkers from the fields of engineering, technology, government, accounting practice and academia. While it’s been a while since that meeting, with discussions over the role of data in the economy continuing in light of the Digital Economy Bill and more, there are some important lessons from that event to consider.

In short, the argument of the RAEng is that data assets are an essential element of the value of UK companies and, with the growth of big data and the internet of things, this is only likely to increase. But currently these assets are not readily recognisable in corporate reporting, nor are they considered in government statistics around GDP. Valuation of these assets in internal management accounting is also rare, and overall the RAEng argues that the current state of affairs undersells UK productivity, doesn’t appropriately encourage the use of open and structured datasets, and under-promotes the need for security around data. By bringing data valuation techniques to the fore, the RAEng argues, the economy and society could benefit.

The first point addressed was the value of data. Most felt that raw data had low value (especially if it was unstructured); the value was only theoretical at that stage. If the data was analysed into understandable information, then the value became clearer. The value is greatest once that information is then applied to reality. Knowledge about the business, its customers or its suppliers is gained as a result. Part of the difficulty in valuing data is that the cost of performing a refinement process – and the value of the knowledge ultimately attained – is largely unknown until the work has been undertaken.

Another key driver of value is synergy. Combining two data sets can lead to fresh perspectives and new knowledge that is not available in either data set separately. This can lead to A+B being greater than A and B separately. Many have commented that this value can be generated by open, but not necessarily free, data – and therefore there is a need for common data standards.

Of course, data loses value with time as its validity and relevance fade, depreciating just as physical assets do, unless it is actively maintained. Auditing and refreshing data is as important as collecting and analysing it. Data is valuable to more than just the company that collects it –

including the criminal element. Cyber security is essential for any modern organisation, but a data-driven one doubly so. Regardless of the quality of its security, organisations holding significant amounts of personal data also open themselves to issues with public trust. People will always be suspicious of an organisation that holds their personal data and so the appearance of security and trustworthiness – and user control over their data – is as important as the security itself.

Valuation methods

Current accounting practice around intangibles is quite limited, with the majority of accounting rules having been designed by 19th- and 20th-century organisations. Currently, the rules would allow for recognition on a transactional basis – ie, if the data could be valued against an external market, or as an element of goodwill on a business combination. Value in use is theoretically relevant, but again would require reliable external yardsticks to establish. Capitalised development expenditure rules are also quite narrow. Most data collection and analysis work is unsuitable due to rules on closeness to generation of identifiable cash flows, and severability from general costs of running a business. Of course, accounting rules have to lead to results which external stakeholders can rely on to make their financial decisions, and the value of data and similar internally generated, intangible assets can be overstated if rules aren’t clear and conservative.

Within the context of national statistics and internal management accounts, there are fewer issues with these measures and greater creativity can be applied. For example, the Office for National Statistics could estimate the value of data by measuring the output of those it identifies as data scientists (formerly known as statisticians) or those in connected roles. However, this method also suffers because as computing power exponentially improves, the ‘cost-per-insight’ in data analysis reduces too. Another suggestion is to use replacement value for data. However, this suffers many of the same problems as the other methods in terms of uncertainty.

In December, I attended a follow-up meeting co-hosted by the RAEng and the Royal Statistical Society. The panel discussion focused on issues such as data quality, identified as a key component of what drives the value of data, but also on the asymmetry between parties in a data economy. By nature, the seller of data is in a better position to know its worth than a buyer – there is an information gap that leads to a relatively inefficient market. This gap is even more apparent between information collectors and subjects – we may not realise just how much a company learns about us when opting in or out of information-sharing arrangements and apps.

Ultimately these considerations have to be worked on. Ignoring the value of data assets isn’t tenable in the long term, any more than ignoring the effect of the internet on business models over the past twenty years was. In my opinion, the first step has to be for regulation and standardisation to make the way for true markets in data. This will create huge synergistic value for the economy. It will also make a market for valuations to mark against, creating a true picture of the value of data-driven companies across many sectors. No matter what approach is taken there will be a need to create transparent permission processes and data ownership for individuals whose data is part of these data sets.

Investment in cyber security research and training to protect that personal data will also be crucial.