Business Consulting

Data steward: who they are, what they do and when you need one

Who is accountable, day by day, for the quality and correct use of an organisation’s data? We find out in this article. 

tasselli incastrati a simboleggiare il lavoro quotidiano del data steward, nella nuova copertina del magazine digitale di Bliss

The data steward is the person tasked with looking after a set of data: ensuring it is correct, described, accessible to those entitled to it and used according to the rules. They do not decide what purposes the data serves, which falls to whoever is accountable for it, but they ensure its integrity every day. The term lives in two worlds that do not talk to each other: business, where the data steward works within data governance (the system of rules and responsibilities for data), and research, where they support those who produce scientific data. This guide keeps the two apart, because the roles resemble each other more in name than in the work.

In brief

The data steward in 4 key points

  1. 1

    The data steward looks after the quality, documentation and use of data within a defined domain.

  2. 2

    It does not decide the purposes of the data: that decision remains with the data owner and, above them, with those who govern the organisation.

  3. 3

    In business and in research the title is the same; the tasks and the data differ.

  4. 4

    A role that is unwritten and without authority becomes a person who reports problems without being able to solve them.

Data steward: meaning and translation

In English, steward denotes someone who holds in custody something that is not theirs and is accountable for it: the administrator, the custodian. That is the essence of the role. The data steward neither owns nor chooses the data: they look after it on behalf of the organisation. In the data management literature, data stewardship is accountability for data and for the processes that ensure its effective control and use (DAMA International, DAMA-DMBOK). In Italian no translation has taken hold: the English term is used, and sometimes “responsabile della cura dei dati” (the person responsible for the care of data), which captures its sense well.

What a data steward does

The work changes with the context, but it revolves around five activities. We have gathered them into our own classification, which combines the guidance of DAMA-DMBOK with that of the research literature cited below.

The five tasks of the data steward

What it looks after, and what happens if no one does.

DutyQuestion it answersIf it is missing
DefineWhat does this data mean, and in what form is it valid?Different departments use the same term with different meanings
Check qualityIs the data correct, complete and up to date?Errors that repeat in every use of the data
DescribeWhere does it originate, who updates it, what is it for?No one knows whether a piece of data is reliable or official
Regulate accessWho can read it, modify it, share it?Data used by people who should not have it, or withheld from those who need it
Report and resolveWho steps in when the data is wrong?Reported issues go unanswered

Bliss classification, combining DAMA-DMBOK and the literature on data stewardship in research.

The duty most often neglected is the last one: reporting a problem is easy; solving it requires someone with the authority to do so. That is why the role does not work on its own, and depends on whoever appoints it.

Data steward, data owner and other roles: who does what

Several roles work on data, and the titles overlap. We distinguish them by type of responsibility: who decides, who curates, who manages the technology, who oversees compliance.

Who does what with data

Four responsibilities that are often confused.

RoleQuestion it answersLimits of the role
Data ownerWhich rules, purposes and access rights apply to this data?Decides, but does not look after the data day to day
Data stewardDoes the data comply with the rules? Is it documented and of good quality?Maintains, but does not decide the purposes
Technical data managerWhere is the data, and how is it technically protected?Manages the systems, not the meaning of the data
DPO (data protection officer)Does the organisation comply with personal data legislation?Oversees: must have no conflicts of interest (GDPR, Art. 38(6))

Bliss classification by type of responsibility. Titles vary from one organisation to another: what counts are the written duties.

The data owner is accountable for a data domain: they define its rules, purposes and access. The data steward applies those rules and checks they hold in the data. What holds the two figures together is data governance; the overall picture is in the Data Governance Framework, while the role of the decision-maker is explored in the article on the data owner.

One distinction matters in law. The data protection officer (DPO) required by the GDPR may carry out other tasks, but the controller must ensure they do not give rise to a conflict of interests (Regulation (EU) 2016/679, Art. 38(6)). The data steward looks after data for the organisation; the DPO monitors compliance with the law: whoever holds both roles at once is checking their own work.

A test

Choose a piece of data that appears across several channels, such as the description of a service, and ask three people from different departments where the official version sits and who can change it. If you get three answers, you are missing an owner and a data steward, and that is where to start.

Our own test, not drawn from sources.

Why a company needs a data steward

For two reasons: one regulatory, one of governance.

The regulatory reason is the quality of personal data. The GDPR requires that it be accurate and, where necessary, kept up to date, and that every reasonable step be taken to erase or rectify inaccurate data (Art. 5(1)(d)). The controller is responsible for compliance with these principles and must be able to demonstrate it (Art. 5(2)). Someone has to see to this every day, and the data steward is the figure to whom the organisation entrusts the task.

The governance reason is accountability. The ISO/IEC 38505-1:2026 standard on the governance of data is addressed to organisations’ governing bodies (owners, directors, partners, executives) and gives them principles for the effective, efficient and acceptable use of data. Data governance is therefore a matter for the top. The data steward is the point at which it becomes day-to-day work, but does not replace it.

The Bliss view: delegating data needs a name

The weak point in many organisations is not the lack of a data steward: it is that no one has established what they are accountable for. This happens when data originates in one department, is used by another and corrected by a third, and each considers it someone else’s. The result shows externally: the same client with two records, the same product with two descriptions, the same message in two versions. For a brand, these are brand inconsistencies before they are filing errors.

This is why we see the data steward as a Brand Governance case: written roles, clear domains, approval levels, someone with the authority to correct. Brand Governance deals with the consistency of the brand over time, and data is part of that consistency.

How to introduce a data steward into a company

Six steps to introduce a data steward

From choosing the data to verification.

StepQuestion it answersWho is accountable
1. Choose the critical dataWhich data do decisions, clients and legal obligations depend on?Management
2. Appoint a data ownerWho has authority over the rules in this domain?Management
3. Designate the stewardsWho already knows this data from day-to-day work?Data owner
4. Write down duties and decisionsWhat can the steward decide alone?Data owner and management
5. Set quality standards and reportingHow is an error reported and corrected?Data steward
6. VerifyAre the rules working? Is the critical data still the right data?Management

Bliss’s own synthesis, not a third-party methodology.

There is no formal pathway in Italy leading to the title: those who hold the role most often have a degree or doctorate in the data’s subject area and, in many cases, IT skills. In business, the steward is generally the person who knows the data domain best, rather than the person who knows the technology best.

Domande frequenti

What does a data steward do?

Looks after a data set on behalf of the organisation: defines its meaning, checks its quality, documents it, controls who may use it and reports or resolves problems. Applies the rules set by the data owner. In research, also supports those producing scientific data, with management plans and FAIR principles.

What does steward mean in Italian?

Steward denotes someone who holds in custody something that is not theirs and is accountable for it: an administrator, a custodian. In Italian the term has no established translation, and the English is used. “Responsabile della cura dei dati” (the person responsible for the care of data) is a faithful paraphrase, because it stresses custody rather than ownership or decision-making.

What is the difference between a data owner and a data steward?

The data owner is accountable for a data domain: they set its rules, purposes and access. The data steward applies those rules and checks every day that the data complies with them. The first decides, the second curates. Without an owner, the steward can only report problems, with no authority to resolve them.

What is data governance?

It is the system of rules, roles and responsibilities through which an organisation decides how data is produced, used and protected. The ISO/IEC 38505-1:2026 standard places it among the responsibilities of governing bodies. The data steward is one of its operational tools: they apply the rules to the data, but do not set them.

What is master data?

This is the core data describing the entities the organisation works with, such as customers, products or suppliers, and which many processes use. Because an error repeats itself everywhere, this is where a data steward is most useful: the same data must have a single official version.

What is the difference between a data steward and a data manager?

In common usage, “data manager” refers to whoever handles data in an operational or technical sense, and the term has no shared meaning. The data steward has a more precise responsibility: the care and quality of the data in a given domain, within the framework of rules set by the data owner. It is better to look at the written duties, not the title.

Is a data steward the same as a data scientist?

No. The data scientist analyses data to produce models or conclusions. The data steward deals with what comes first: making sure data is correct, documented and usable. The two roles work together, because analysis built on uncurated data yields unreliable results, but their skills and responsibilities differ.

Can the data steward also be the DPO?

It is inadvisable. The GDPR allows the data protection officer to perform other tasks, but the organisation must ensure they do not give rise to a conflict of interest (Art. 38(6)). The data steward looks after data for the organisation, the DPO oversees: whoever holds both roles is checking their own work.

How do you become a data steward?

In Italy there is no single formal pathway. In research, surveys show that most hold a degree or doctorate in the relevant data discipline and, often, IT skills. In companies, people reach the role by knowing the data domain well and learning the governance rules. According to the University of Palermo, there are dedicated master's programmes in Italy and France, as well as summer schools.

What does a data steward do in research?

They help researchers manage data across its life cycle: data management plans, description with metadata, application of the FAIR principles, deposit in trusted repositories and sharing in line with ethical and legal rules. In the Italian survey, the most frequent activities are support for the FAIR principles (34 responses out of 38) and for management plans (31).

How many data stewards does an organisation need?

It depends on the data and the structure: there is no benchmark figure. In the Italian survey on research, 24 of 38 respondents report two to ten people with equivalent functions, and 7 a single person. For a company, the number of critical data domains matters more than headcount: each domain needs a steward with an owner.

Fonti e riferimenti
  1. DAMA International, DAMA-DMBOK (Data Management Body of Knowledge)
  2. Giulia Caldoni e altri, Data stewardship in Italy: Findings from the first national survey, Open Research Europe, 2026, 6:45
  3. Comunità Italiana Data Steward, Manifesto della Comunità Italiana Data Steward (CIDS), 31 marzo 2025
  4. Università degli Studi di Palermo, La figura del data steward
  5. ISO/IEC, ISO/IEC 38505-1:2026, Governance of data
  6. Parlamento europeo e Consiglio, Regolamento (UE) 2016/679 (GDPR), artt. 5 e 38
BLISS®© 2026. ALL RIGHTS RESERVED
Scritto il
Contact