Photo by Suzy Hazelwood from Pexels

It is time to consider ‘the value of data’ – a number of recent conversations (such as with Stuart Ravens and Martin Paver), an excellent recent article on Highways England’s data valuation and related events such as the emergence of a proposed international standard to develop a ‘data value index’ show that this is a timely subject.

We are increasingly (and correctly) being reminded to ‘treat data like an asset’, in other words to recognise that it is something of value to an organisation and should be managed and nurtured, but also that this value can be degraded if the asset is not cared for.

Related to this is a growing desire to understand the absolute value of the data asset.

Why value data?

There can be a number of reasons why an organisation may want to obtain a value for their data that include:

  • To support sale of the organisation or a stock market listing
  • Acting as security to obtain external financing for the organisation
  • To price data services provided to customers
  • To evaluate the cost (and potential value) of data being purchased/ acquired by an organisation i.e. how might it increase the valuation of the organisation?
  • To raise staff awareness about why data is important to the organisation and to encourage better data practices

Types of data valuation

An important consideration is the ‘type’ of data valuation that you are trying to obtain. The two extremes could be typified as:

  • An accounting valuation – a value that follows a defined methodology, could be repeatably obtained by different assessors and could support other financial processes, such as organisational valuation or backing for external finance
  • An indicative valuation – a value derived from a simpler methodology that may be difficult to repeat (particularly if different parties are involved) and is unlikely to be robust enough to support balance sheet valuations and external financing options

The challenges of valuing data

Understanding the value of a second hand car is relatively easy – they come in limited variants and there are many examples of comparable car sales to allow a reasonably accurate assessment of value to be obtained based on condition, age, mileage etc.

Valuing data is far more complex, in part due to the complex nature of data in organisations and the many uses it can serve. Data valuation can depend on a number of factors:

  • The benefits it can directly provide e.g. services to customers, decision making to support those services N.B. different organisations with identical data may correctly have different valuations due to the different benefits that the data will provide to each organisation
  • Potential benefits when combined with other data e.g. combining your existing customer data with newly acquired demographic data
  • Potential future benefits e.g. information on how to respond to an emergency or how to safely decommission a nuclear power station only has value when those events arise in the future
  • Market value if sold e.g. aerial survey data you have gathered that could be purchased by farmers to calculate fertilizer dosing requirements
  • Costs of not having the data e.g. costs avoided by not having to survey customers to understand the number of people in each household
  • Intellectual property, designs, trade secrets may be stored as data/ information and allow an organisation to maintain a competitive advantage

Organisational context will also have a large impact:

  • The costs or ease of gathering the data again, for example, data on buried cables and pipes can only be obtained when the pipe/cable is being laid or if you dig down to it; data on assets in a nuclear power station can only be gathered when expensive and time consuming safety procedures have been followed
  • Information on steel reinforcement in a concrete structure, if it did not exist or had been lost, could only be gathered if you were to demolish the structure
  • Being unable to ‘rewind’ time to determine the batch or serial number of all components used in a car – for example to be able to identify vehicles requiring a safety recall
  • Contract details for a mobile phone customer who signed up four years ago – what did they actually sign up for if you cannot access their contract?

Factors that also are crucial to understanding value:

  • The quality of the data – just because you have lots of data, does not mean it is accurate or usable!
  • Sector specific differences, for example, customer data in retail is very different to healthcare data
  • The level of rigour required – is this a value solely for informational purposes, or does it have to support, for example, a corporate valuation and therefore need a high level of accountancy rigour

How (not?) to value data

One way to value your data asset would be to:

  • Undertake a full data and process inventory for your organisation
  • For each data set and process: evaluate the benefits it supports; the possible future benefits it could provide; the potential revenue that could be earned if the data were to be sold; potential value if the data were combined with other data sets; and the cost to recreate this data set
  • Repeat across the whole organisation
  • Undertake a formal process of converting all these benefits and values into a formal data value
  • Ensure that the methodology for the above is documented and there is a full audit trail to ensure any future audits of data value have a clear understanding of the derivation of the data value figure

This process would almost certainly take significant time and effort to complete and would therefore be a very expensive process to undertake for your organisation. The recent work by Highways England that valued their data at £60 billion clearly took a large amount of work (at an unstated, but probably large cost) but due to the size of this valuation will have helped justify its cost. It is also interesting to compare the £60 billion data valuation to the £300 billion corporate valuation i.e. data is 20% of the value of the organisation. It is likely that the data of other asset intensive organisations will be a similar proportion of their total valuation.

Clearly, the context of other organisations will be very different and would preclude such valuation approaches. Additionally, as there are few definitive approaches to valuing data, it will be very likely prove challenging to gain approval for a similar project in other organisations. The lack of a clear valuation methodology means that an audit of the valuation may produce different results potentially discrediting the valuations obtained.

Is there a benefit to attempting to deliver an accurate data valuation? Perhaps if the valuation is used to support a sale of the organisation, a stock market listing or is used to secure finance, then there may be a tangible benefit. However, it is unclear whether there will be clear tangible benefits for many other situations.

A more pragmatic approach

I am a strong believer in adopting pragmatic solutions, so perhaps a more pragmatic approach is to not value your data!

If you assume that your data has a value of £X and that this is a large and unspecified figure (using the Highways England example, this could be a significant percentage of your corporate valuation), then you can perhaps think about how this value X will change if:

  • If you are able to create new products and services using your data that are worth £A (which should be straight forward to calculate) then the value of your data is now £X+A
  • If you identify process inefficiencies using the data with an annualised impact of £B, then removal of these avoidable costs will increase the value of the data to £X+B. Again, it should be straightforward to calculate B
  • If you identify poor decision making caused by poor or missing data with an annualised impact of £C, then these are avoidable costs that will increase the value of the data to £X+C. Again, it should be straightforward to calculate C (although this may be reduced by any costs of improving the data)
  • If you identify that some of your site survey data can be sold to third parties, then this sales revenue £D could be viewed as increasing the value of your data to £X+D
  • If you use improved data governance to reduce the number of data stores (by combining some) then the cost of data migration for future projects could be reduced by £E (along with a reduction in the risk associated with data migration), so again the data value could be viewed as £X+E
  • Etc.

Therefore, if you improve how you manage and exploit your data, your overall data valuation could have increased to £X+A+B+C+D+E+….

If you don’t worry about the absolute value of X, then you can still demonstrate how you have improved the data value of the organisation by listing the incremental benefits £A, B, C, D, E, etc.


There are three basic choices here:

  1. Spend a lot of time and effort to gain an accurate valuation £X
  2. Spend time identifying and delivering £A+B+C+D+E+… without worrying unduly about the value X
  3. Carry on as you are

As a data professional, option 3 is not a comfortable place to be.

Option 1 is expensive and time consuming, but may be the correct approach for your organisation depending on its particular circumstances and requirements.

Option 2 is probably the most pragmatic one (that also more naturally leads to improvements) by identifying and delivering incremental improvements

So which approach to data value will you choose?

Tagged on:         

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.