data stewards
data stewards

Data steward, a new profession that has been emerging over the past three years. Learn about this new role deployed in a range of business sectors. By Pierre-Sylvain ROOS, Chief Technical Marketing Officer.

a new job ?

It all started just under three years ago when, for the first time in France, customers started looking for Data Steward consultants. Since then, requests for Data Stewardship services have continued to grow. They are now common and concern more and more business sectors. 

Obviously, at the beginning it was necessary to understand what the slightly odd term of Data Steward actually meant. Everyone knows that a steward is the person in charge of welcoming and serving passengers on a plane. So, why would data need a steward?

The first part of the answer: data is growing and the volume of data produced in the world is growing exponentially and constantly.

The second part of the answer: this mass-produced data, or Big Data, can be used by companies to add value. To do this, they need to know what data to use and in what context. Therefore, all this data has to be managed in terms of lifecycle, quality and meaning.

That is why since the early 2000s, large companies started putting in place special systems to structure and improve the management and security of their data.  

In particular, major international banks have been at the forefront of developing Data Governance with a notable acceleration in 2013 following the new BCBS 239 standard entitled “Principles for effective risk data aggregation and risk reporting” in the framework of Basel III regulations.

As for Data Stewards, it was the Data Governance Institute (founded in 2003) that defined their roles and functions as key players in a mature Data Governance organisation. Data Stewards take care of Data assets. It is a profession in its own right and has been fully defined, structured and described since 2014.

And in France? Well, when I was recently browsing a book published by Editions de l’Etudiant about Data jobs, there was a complete absence of any mention of Data Stewards - not a trace or line anywhere.

Ditto if you look for a Data Steward in a job search - candidates are few and far between! 

In France the job of Data Steward is basically unknown, let alone recognised. Perhaps the issue lies with the name, many Data Specialists might actually be Data Stewards without knowing it!  Moreover, by digging a little deeper, we can find corresponding terminology: data guardian, data quartermaster, data custodian, data coordinator, data administrator, etc.   

So, if you want to find a Data Steward, it’s best to focus on the function and the role rather than the job title in order to have a minimum understanding of Data Stewardship.    

there are several types of data stewards

This is the first surprise we come across when we start learning about this job. There is not ‘one’ but ‘several’ types of Data Stewards each with complementary roles and functions, more or less defined, depending on the size and maturity of the existing Data Governance organisation. In some instances, there is even a Chief Data Steward (who may also be called a Data Governance Manager).

They are responsible for implementing the operational aspects of data governance and carrying out the day-to-day work necessary to ensure that the data is available to the entire organisation. 

Almost all sources define five or six types of Data Stewards (see below):

  • Entreprise Data Steward: he/she works in a cross-cutting manner in the entire company and, usually, is responsible for coordinating Business Data Stewards;
  • Business Data Steward: he/she is in charge of corporate business data, notably managing the lifecycle and quality of data. He/she facilitates data understanding and control by the business line;
  • Domain Data Steward: this is the same as a Business Data Steward but has a more specific role i.e. he/she only deals with data that is shared by several business lines, and plays a decision-making and coordination role; 
  • Project Data Steward: he/she works on projects and ensures alignment between the used data and the project’s needs;
  • Technical Data Steward: he/she provides technical expertise on the systems used to produce, transform and store data; 
  • Operational Data Steward: in particular, he/she assists the Business Data Steward and works directly with the data. He/she ensures that data operations are valid and effective.

All these different types of Data Stewards work together under the authority of a Data Governance Programme Office (DGPO); the entity in charge of Data Stewards within the DGPO, which is called the Data Stewardship Council.

Obviously, depending on organisational choices and the maturity of data governance, there will be more or less focus on certain aspects of Data Stewardship. Currently, you will not find all these types of Data Stewards in the same organisation. 

a key objective improving data quality

Data Stewardship is a development of Data Quality professions. 

Data Stewards not only focus on the intrinsic quality of data but also its semantics and usage.

Therefore, they are responsible for managing traditional qualitative aspects such as data formats, technical validity, completeness, integrity and uniqueness.

But they also have to guarantee the conditions under which the data can be used, draw up and document usage rules, and ensure that the data is meaningful, accurate and unbiased when analysed.

For example, here are three typical issues that a Data Steward might have to address (with an example for each):

  • the data is unusable because its source is unreliable: this happens due to smartphones’ internal date and time functions. The best example of this well-known problem is the success of the game Candy Crush Saga. To earn extra lives without waiting or paying, the most avid fans (i.e. almost all) forward the date and time on their device. Candy Crush Saga has over 380 million active players i.e. around 16% of all smartphone users in the world. Therefore, it is necessary to use an external clock to timestamp a smartphone event.
  • Certain data have limited validity in terms of time: the size of a child or teenager changes significantly even very significantly until he/she reaches adult height. To make sense, this data must always be associated with a date. 
  • The data is correct in its form but it is not plausible: if a medication delivery date (at a pharmacy) is prior to the date on the prescription then at least one of the two dates must be wrong. The two dates must be checked and corrected.

In each situation, the Data Steward is responsible for defining the conditions and restrictions that govern the use of data, as well as negotiating data management decisions with stakeholders.

In addition to this, one of the Data Steward’s key concerns is to prevent the Data quality deteriorating. To do this, he/she must put in place metrics to continually monitor the data.

Finally, no data without metadata. Metadata is used to define and describe data; Data Stewards are particularly interested in this. It is used to manage and qualify data, and is a valuable source of analysis and information in and of itself.  

In all cases, the Data Steward’s main role is to proactively seek out any data quality issues as early as possible i.e. before data users are impacted. As a result, he/she often works directly with data producers and ensures that any detected problems are corrected.

The most obvious benefit of Data Stewardship is the visible and continuous improvement in data quality. 

data steward tools

A mature Data Governance organisation is a properly equipped (i.e. with tools) organisation. The efficiency of the Data Steward is absolutely dependent on the efficiency and effectiveness of the available tools. Tools are also the most identifiable marker for defining the level of development of Data Stewardship within a company.

Four key tools are particularly important for Data Stewards:

  • Data Stewardship portal: this is a communication tool used to share the work performed by Data Stewards and their results with stakeholders in the company (management policies, procedures, contacts, etc.);   
  • Data Stewardship Wiki: this contains the specialised terms, methods and best practices used by Data Stewards; 
  • business glossary: this contains data nomenclature, definitions of business terms, usage and validation rules, who is the data owner and what is their security classification, as well as any associated quality requirements;
  • metadata repository: this system references and describes all the physical metadata (logs, timestamps, size, etc.) and techniques (file structure, interface format, etc.) plus any links.

And in terms of operational activities, there must be a ticketing system to log, document and address any data problems.

we are not born data stewards, we become them…

As you can see, we have a long way to go before we find Data Stewards in all companies. To date, the most advanced Data Governance organisations are themselves undergoing transformation. This includes major banks, insurance companies and, recently, telecom operators and energy companies.

The development of Data Governance and Data Steward jobs often goes hand in hand with the development and industrialisation of Big Data activities. All these corporations are moving towards Data-driven business models i.e. where Data is the main asset and focus.

This is why it is interesting to note that there is a Data Stewardship maturity model, which is a particularly useful and effective support tool to guide and orient these companies in their Data Governance deployment and structuring processes.

It relies on four pillars (or four key questions):

  • has my organisation integrated Data Stewardship and how does it support it?
  • are the roles of Data Stewards well defined and how are their activities structured?
  • is Data Stewardship based on properly formalised and implemented standards, policies and processes? 
  • is the organisation able to measure and prove the value created by Data Stewardship?

The level of response given to each of these questions makes it possible to work out what to do next in stages.

If we go back three years, most companies that had implemented Data Governance in France had not properly identified or formalised the role of Data Stewards. This has now been done and the recent roll-out of the General Data Protection Regulation (GDPR) will be an additional lever for the increase in Data Stewards.

Many current Data Specialists, regardless of their job title, have some way to go before their role as Data Steward is fully recognised and valued. However, there is no doubt that this movement has already begun.

So to conclude, and to refer back to the initial question posed in this article - if you are still asking the question or wondering where all the Stewards are - the answer is actually simple: they’re on their way!