The chief data officer is an emerging position tasked with overseeing projects that aim to derive value from enterprise data sets: some observers think that CDOs could assume a defining role in shaping corporate success for the next 10 years.
As data is increasingly perceived as a dynamic asset, more thought will be given to how best to manage it – not only in terms of storage, availability, quality and analytics, but also the architecture that allows for data being accessible and ‘malleable’.
Bigger data management covers IT, business strategy, statistical analysis and more. Accordingly, the job title of chief data officer (CDO) is becoming more prevalent in executive hierarchies, although some estimates suggest that there are now only 150 CDOs in organisations around the world.
However, marketwatcher Gartner has recently predicted that by January 2015, 25 per cent of large global organisations will have appointed CDOs. Gartner also reckons that some 65 per cent of current CDOs work in the US, while 20 per cent are UKbased; there are, however, CDOs in over 12 countries now, the analyst reports. In addition, over 25 per cent of CDOs are women – that’s almost twice as high as the proportion of chief information officers (CIOs). This is most common in heavily regulated industries, media and government, Gartner found.
According to a recent report in the Financial Times, the speed with which organsiations are appointing CDOs “will not only define the shape of business, but also the next generation of men and women who will lead them”.
Standard specifications of a typical CDO’s roles and responsibilities may differ from enterprise to enterprise. According to one standard definition, CDOs are responsible for enterprisewide governance and use of information as an asset, via data processing, analysis, data mining, information trading and other means. They won’t necessarily report to an established CIO. Stuart Coleman, commercial director at The Open Data Institute (the business facilitation body founded by Sir Tim BernersLee and Professor Sir Nigel Shadbolt), agrees that the CDO “is essentially responsible for determining how data can be used across an organisation and the operational environment to drive better business outcomes.”
In a significant number of corporate entities a range of data has been (and is being) collected and stored, but has not really yet been fully assessed and interrogated for strategic gain. More companies starting to use data analytics tools across their business operations means that the CDO will play a pretty vital role in ensuring that data can make a meaningful contribution to informed business strategy. The standard appointment of a CDO in any organisation that values its ‘data estate’ would appear a necessity.
“Historically, data came within the remit of the IT department, and was seen as something to be stored – what has changed is that data is now an active asset that can have a lot of value extracted from it… It is therefore a business issue,” says Laurie Miles, head of analytics at business analytics software and services supplier SAS. “For some companies like Google, for example, the business is very datacentric. They treat this very seriously, and for them the CDO is a natural and important role. Other companies are further behind the curve on this.”
Appreciating the value of data
A sector that seems to be more appreciative of the value of data is banking and finance: it was shaken by lack of data visibility and transparency during the 2008 crisis. The retail sector too has recognised that it is not just the transactional data, but all the information around that which can inform future campaigns and distribution channels.
The telecommunications industry has also longidentified the value in its data – here it is not just the static data that is attractive but more the data on behaviour, social networks and suchlike that is ripe for harvesting. Customer billing data, for instance, not only provides insight into usage patterns that can help forward planning, but also provides evidence of trends that could inspire new revenuegenerating services and products. More recently the UK National Health Service (NHS) and central government are taking strides – sometimes controversially – to improve the quality of their centrallyheld data, and to harness it to inform future strategy and make better informed and supported predictions.
“No matter what the industry, the trend is for there to be wide recognition of the benefit of data and that it is an asset,” acknowledges Chris Nott, CTO big data & analytics, IBM UK & Ireland. “The importance of deriving analytical insight from data has never been greater.” In practice this means that data needs to be able to look at the past, the present and then, on the basis of those, be able to predict the future. An early example of this, says Professor Jim Austin, CEO at data analytics company Cybula, was looking at data sets from aircraft taking off and using that to inform a maintenance programme for aero engine manufacturer RollsRoyce.
‘Cleansing’ and preparing data to allow for a single view allows for it to be analysed not just for a single purpose, but in any number of areas. According to Nott, by applying analytics in one area – to identify fraud, say – there is an overall business benefit which then translates through to others such as marketing, profitability, inventory, etc, also being able to use the same data sets.
“The CDO making the right data in the right place at the right time gives an organisation the opportunity to innovate, and to harness data for better decision making and strategic policy,” Nott says. “It also gives better overall visibility and harmonisation – and standardisation.”
Indeed, creating that usable data set in the first place all comes down to how data is governed over the whole organisation. At Barclays bank, for example, firmwide governance of data is something that is being looked with a view to simplification and gaining both efficiencies as well as a better overall quality of data.
In the NHS, meanwhile, information governance, which involves ensuring that the necessary safeguards are in place across the organisation for protecting and sharing information appropriately, is higherup on the priority list. There is also a focus on information standards. The first two, thus, inform strategic intelligence and transparency – which, in turn, improve internal accountability. Accountability and regulation go handinhand with good data governance and transparency.
The finance industry, one of the most heavilyregulated industries, was in fact one of the first to get behind the concept of a CDO because it needed to clean up its data provenance. “Banking is under intense scrutiny and pressure to change for the better,” says Barclays Bank CDO Dr Usama Fayyad, who was said to be the first person appointed to the role of CDO in his previous role at Yahoo. “One of the benefits of centralising data is that it can become easier to comply with regulatory changes. For example, if a business within Barclays is storing the same data in three or four places – and we move it into one central database.”
The data traction attraction
There are inherent advantages to having someone look after both the way in which critical enterprise data is prepared and stored, as well as the way it is then used for future decision making and strategy. IBM’s Nott describes the CDO as a critical lynchpin. “It’s become apparent that someone senior with an overview was missing,” he believes, “and putting into place the right expertise and mechanism to bring all that data together and make it work is now key.”
In practical terms this means that the CDO should serve as an executive interface between the various business branches – departments, directorates, project groups – and touch points that data has. It means working with the IT department to create appropriate systems architecture and the right environment to start working on the data analytics.
This aspect of the process will be scheduled in conjunction with a ‘data scientist’ or even what could be called an ‘enterprise statistician’, who would not necessarily be formally attached to the traditional IT function.
The Open Data Institute’s Coleman says that key to the skillset of the enterprisewide CDO is being able to manage key relationships with all aspects of the enterprise from legal to marketing to compliance. “Marrying domain knowledge of your industry with a broad range of data experience and skills is critical to be able to know what questions you can – and should – ask of your data to make a difference to your business,” he explains.
CDOs are “most commonly from an IT background, but in the future the CDO probably more likely to come from a business background”, agrees Miles at SAS. “They need to be able to identify a business problem, and then try to work out what data can do to solve it.”
He also thinks that Master of Business Administration (MBA) degrees should have a quantative analyst element incorporated into them – this would do much to help weave the data and the business knowledge together, he claims. Indeed, the challenge of having a CDO at anything other than enterprisewide level is that there is a lack of overall insight. In an age of centrallyheld single ‘golden’ copies of data, having anything other than an enterprisewide CDO can start to seem a bit nonsensical.
Fitting the CDO into business
Barclays’ Fayyad says: “I view the CDO job as a mix between business and technical responsibilities. An effective CDO, in my opinion, must drive some business strategies as well as own technical execution and delivery with an emphasis on excellence in data management and access: both by humans (insights) and machines (systems).”
The way in which the CDO works with other datarelated professional specialisms also varies. The Big Data vogue is causing many businesses and other organsations to rethink their strategic attitude toward their data ‘estates’, and part of that exercise is to look at ways to apply different forms of value to that data. Another aspect is to examine ways in which data generated and ‘owned’ by one part of the business could possibly be exploited by a different part of the business. There is potential for valueadded crossfertilisation, but some of the protocols on which data management has for years been based will have to be rethought.
“Data is owned and used by the business overall, but managed by IT,” says Jane Griffin, executive advisor at Deloitte Analytics Canada. “Within IT, however, a data manager looks at how the architecture can best serve the data – can it be consolidated? Can a different application be used?
What are the IT governance roles to be applied around the data?”
Bringing the data to the table is then the role of IT. Being able to gain insight from it is the role of the data scientist or statistician who commonly sits within a business unit. “Directed by the CDO, the data scientists look at the data itself, and make it meaningful by devising the methodology and the modelling around it,” says the Open Data Institute’s head of statistics Ulrich Atz. “The CDO sets the direction, and data scientists provide the insight. They can be from a social science statistical world, where they are looking to find cause and prove a trend, or they can be from an IT background, a computer scientist or software engineer – and thus have strength in the architecture.”
Indeed, although data scientists do not necessarily have to have a background in IT, they are apt to be highly technical people. A deep knowledge in computer science or systems engineering is pretty essential in the background of an effective data scientist. And although statistics is a technical field, data scientists who are statisticians by training still have to have strong computational skills.
Interest in Big Data is no longer confined to corporate strategy documents. Soon stakeholders in the wider world – institutional shareholders, say – will start to expect a Big Data strategy to deliver value to the balance sheet. The number of CDOs looks set to increase as organisations realise the extent to which their data estates could be exploited for strategic advantage.
Future CDOs could come from IT, statistical, or computational backgrounds; but they also need to have the business skills to define how to make best use of data and drive data insight forward. Marrying technical and the business element together is rarely easily achieved, but is clearly required if organisations are going to successfully define and create the architecture required to derive insight.
Verticals putting the squeeze on their data
Usama Fayyad’s role as Barclays’ CDO was created because the bank wanted to “look at data strategically to create new insights… for customers and clients both internally and externally”, he says. Within the banking industry the need to hold data centrally and to have reliable data provenance is in part a reaction to the demands of the respective regulators. This has since spread to the business, where the value in having centralised data and being able to use it to identify possible fraudulent activity, engage in targeted marketing and crossselling has assumed much value.
Part of Fayyad’s remit is to establish a centralised team called Data Systems and Insights (DSI). “Barclays has a number of separate businesses that have been doing data management in different ways using different technologies,” he explains. “So there is a great opportunity now to build a best in class data function.”
Among retailers, names such as John Lewis are investing significantly in data and architecture to be able to operate across any number of sales channels and bettermanage operations (see E&T June 2014). John Lewis now holds data centrally on a single, thirdparty platform, says its head of IT architecture, Julian Burnett: this has forced a single product data reference, which means greater visibility from sourcing right through to returns.
Other retailers to have invested in this area include Matalan and Argos.
Telecoms companies initially invested in data management to understand customer churn, but are now looking to gain insight into customer behaviour via social networks as a means to provide valueadded services such as targeted offers or disc ounts.
Toward this end, in 2012 the Weve consortium, comprised of EE, O2 (Telefónica UK) and Vodafone UK and covering over 80 per cent of UK customers, was formed to accelerate UK development of mobile marketing and wallet services. It should also facilitate technological standardisation.
Lastly, the NHS, along with some other government departments, is having a good stab at data management. One of the larger projects is the NHS looking at all the various ethnic groups in the UK to try and create a range of scenarios to influence the time it takes to bring new drugs to market.
Dr Geraint Lewis, chief data officer, patients and information NHS England, is tasked with securing highquality care for “the all, now and for future generations”. His team looks after four broad areas: information governance, information standards, strategic intelligence, and open data and transparency. “The only way we can be sure that we are achieving this aim is by collecting and analysing highquality information about the care being provided,” he says.
Author: Alison Ebbage