Part 1 of this blog questioned whether data has value and gave some proof points. Here I discuss why and how organisations should put value on data.
Why put value on data ?
What “raw” data, when combined, constitutes your company’s Intellectual Property ?
If you have some idea of the value of that ( proportional to your company’s market cap ? ) then have you divided and assigned that value to each constituent data set accordingly ? Do you continue to re-assess the importance of each data set as your company grows and its market value increases ?
Clearly, as has been exemplified in part 1 of this blog, data has current value for your business’ operation. But it also holds potential future value … data may reveal value in ways you did not expect.
Even if the data does not directly benefit your company, it may prove valuable to others. Many companies are monetising data sets which, for them, are the digital by-product of their day-to-day operations but for other companies can provide transformational insight and opportunity. In some cases, large enterprises that understand the potential value of their data but realise corporate inertia may delay acting upon it, take the bold step to forge new, culturally separate and more agile subsidiaries to fully capitalise on the opportunity.
But you need to remain ruthlessly pragmatic.
The tendency in most organisations’ IT departments has been to adopt a “one size fits all” approach. Data has generally been recognised as important for the business with standard solutions, processes, practices and controls being used to manage all of it in the same manner. But IT budgets are under constant pressure and, as you may have noticed, data is getting BIG.
Most cannot afford to maintain big data in the same way as their system of record / transactional data.
For example, do the server nodes in a Hadoop cluster adopt the full corporate IT standards by default ? Do they need to have all of the system monitoring and alert agents plus state-of-the-art security that would be installed on your more traditional business systems’ infrastructure ?
The answer may depend on the value you place upon your big data – whether that is from analytical insights currently in operational use by the business, or some expected future insights.
But realise that when the data grows to the extent you have hundreds possibly thousands of nodes, these “ancillary” IT costs may come under scrutiny.
Is it necessary ( or even possible ) to back up your big data ? Do you need to keep all that social media data ? Is the value purely transient with no need to go back to “yesterday’s news” ? For the Internet of Things, what about the continuous terabytes of data generated by sensors on some complex machinery – a jet engine for example ? Preserving full history here may be vital to future product design, warranty, predictive maintenance and, worst case, some failure investigation.
What Disaster Recovery capability do you need for the different kinds of data ? How long can your business survive without that data before it is realised that missing those insights was actually business critical ?
At the very least it will be necessary to evaluate the proportionate cost of storage, and required speed of access / retrieval, according to the business value of data. This leads many companies to a tiered storage approach, so-called “hot, warm and cold” data.
The challenge in the current business environment is that new insights are always being sought and this dictates seamless and intuitive access to join the hot data ( “what is happening now” ) with the cool data ( “what happened over time with what trends” ) to predict “what should I do next”. Data that was once considered “archive” may now be very important for day-to-day business.
This raises another important point – regulation.
Regulations may require you to keep certain data for specific periods of time, handle it in a certain way, and demonstrate various controls over that data. Of course this may vary by industry. But industry lines are blurring, regulation is always changing, and the data you presumed would be exempt from regulation may one day be thought of differently.
With specific regard to Personal Identifiable Information ( PII ), a data set that you previously regarded as anonymised, when combined with other data and enough analytical horsepower, may suddenly yield highly specific insight on an individual. Whereupon that data set takes on a whole new relevance in the eyes of regulators and you may be required to direct your IT spend to manage it accordingly.
How to put value on data
There are no easy answers and, currently, no formal accounting standards that give merit to this new commodity.
Many established IT departments will be familiar with the “CIA ratings model”. In this context that means Confidentiality, Integrity and Availability for which ratings for each are applied to data, or to a system that hosts that data, as a means to assess the extent of security and resilience practices that must be applied to ensure that level of service to the business. It would seem this may be a useful existing framework from which to extrapolate data value.
In a big data context, some are looking at the “4V’s” to help approximate value of data:
- Volume – more data and deeper analysis across current and historical behaviour equates to better decisions
- Velocity – the faster you get the data and can extract insights equates to more timely decisions and response
- Variety – the ability to process all kinds of data equates to a more informed decision
- Veracity – confidence in the quality and consistency of the data equates to the right decisions
It is important to know which parts of the business are using ( relying on ) what data.
For an IT department in any large organisation this can be an arduous task. “Shadow databases” and tooling are emerging in more departments across the organisation ostensibly to support isolated and specific needs. Even if that happens with the knowledge of the CIO, the convenience and popularity will often grow unknown to them or outside their control. The isolated need becomes a corporate necessity and/or the shadow database gets leveraged for other diverse purposes. The value of the data may have increased many-fold but replication, duplication and loss oflineage means no-one is really sure by how much.
More companies are appointing a Chief Data Officer ( CDO ) to help address all these issues. On the one hand, to champion the value of data and both its current and future benefit to the organisation, on the other hand to ensure it is governed properly according to that value.
For now it is probably enough just to start the discussion about the value of data within your organisation. It needs to be a serious discussion with all areas of the business, since it is clear they all contribute and they all can benefit.
Even though the exercise may appear intangible at first, you should find ways that are meaningful to your organisation, that help quantify how data gets used, the value it provides and the corresponding costs to manage it appropriately.
Once the discussion has started, keep it going ! Whether this be the job of the CIO or the CDO ( should your company have one ) it is crucial that the value of data be continuously re-assessed across all lines of business and the organisation as a whole.
Business leaders take great pride in their company DNA. The talent in their workforce, accounted for individually by salary and wages, but collectively worth so much more, defines not only what the company does but how it does it and what it can achieve in the future.
It may be many years, if ever, before data gets recognised as a balance sheet asset by international accounting standards
However, that should not stop you from recognising its true worth to your company. Data may yet turn out to be your star performer.
Follow me on twitter: @robpsap