Independent IFS Cloud practice · Data mesh series, step 8
Data mesh culture: the roles and skills that make domain ownership of IFS Cloud data stick
A data mesh works only when business domain leaders accept that they own their data as a product, and when central IT accepts that its job becomes running the platform. This step covers the role changes, the product-thinking skills to train and the order in which to scale the change.
Key takeaways
- The test of a cultural shift is who fixes a wrong number: in a data mesh it is the domain that produced the data, not the central data team that reported it.
- Product thinking means a domain publishes data with named consumers, written definitions and service levels for freshness and accuracy, and treats feedback from consumers as work to plan.
- The central data or IT team becomes a platform team. It builds the shared tooling and enforces the agreed rules, and it stops owning the content.
- Federated governance replaces a central approval function with a council of domain representatives that agrees a small set of global standards, while execution stays in the domains.
- The change scales from one domain that proves it works, through internal examples and through ownership written into role descriptions, not through a company-wide announcement.
1.Why does data mesh depend on culture more than on technology?
Data mesh depends on culture because its first principle, domain ownership of data, is a change in who is accountable, and no tool can make a manager accept accountability. Zhamak Dehghani lists domain-oriented decentralized data ownership and architecture as the first of four principles, ahead of the platform and governance that support it.
In a central model, business users enter transactions in IFS Cloud and a central team cleans and reports the result. When a figure in a finance report is wrong, the central team spends its time tracing an entry error made in a business area it does not run. In a data mesh the domain that made the entry owns the data product, checks it before publishing and answers the question. The data mesh overview places this as the last of eight steps because domain teams can be trained in product thinking only after domains, teams and products exist from the earlier steps.
The phrase that marks the shift is a move from “the system will fix it” to “we own the accuracy”.
2.Which roles change in a data mesh?
Five roles change in a data mesh, and each change should be written into a role description rather than left to goodwill.
| Role | Before | In a data mesh |
|---|---|---|
| Domain leader | Consumes reports and escalates errors to IT | Accountable for the domain’s data products and their quality targets |
| Data product owner | Did not exist | Decides what the product contains, who its consumers are and which service levels it promises |
| Domain team and key users | Enter data in IFS Cloud and wait for reports | Build and maintain the domain’s products, including their documentation, as described in forming cross-functional data product teams |
| Central data or IT team | Owns the pipelines and reports for everyone | Runs the self-serve platform and enforces the agreed rules, without owning content |
| Governance council | Central approval of every change | Domain representatives agree global standards, the platform enforces them |
Domain leaders take on new work with ownership: their teams check, document and support data that the central team used to handle, and that work starts before any benefit arrives. Budget and time for that work belong in the plan from the start.
3.What is product thinking, and what do you train?
Product thinking is the habit of treating a dataset like something sold to a customer: it has users, a specification, quality commitments and a backlog of improvements. Training for data mesh therefore centres on these product skills.
- Knowing the consumers: the owner can name the teams that use the product and what decisions they make with it.
- Writing definitions: every field has a business definition a reader outside the domain understands, linked to the shared glossary.
- Setting service levels: refresh frequency and an accuracy commitment that are written down and measured.
- Documenting for discovery: the catalog entry is part of the product, as covered in data discoverability.
- Handling feedback: consumer requests go into the product backlog and get a decision, not an informal promise.
- Managing change: a change that breaks consumers is announced and versioned, never applied silently.
The skill set behind the documentation item is metadata management. A training plan for product owners, with the topics in sequence, is in train teams on data product ownership.
4.How do you check whether the organisation is ready?
Organisational readiness for data mesh is checked before any technical restructuring, with questions that can be answered yes or no for each candidate domain.
- Is there a named leader who accepts ownership? A leader who sees the role as an IT task is not ready, whatever the tooling.
- Are there consumers outside the domain? A product without consumers outside its own domain gives no early proof.
- Does the team have time allocated? Ownership work done on top of a full day job stops at the first busy month.
- Are objectives aligned? If the leader’s objectives reward only operational output, data quality will lose every trade-off.
- Is there a platform to build on? Without shared tooling, each domain builds its own and the result does not combine.
A domain that answers yes to all five is a candidate for the first wave. The others wait, and the reason is written down so the decision can be revisited.
5.How does federated governance keep ownership from becoming chaos?
Federated governance keeps decentralized ownership coherent by having domain representatives agree a small set of global standards, such as ISO currency codes or one format for customer and supplier identifiers, which every domain then applies on its own. The council sets rules, and the domains execute them. How to set up that council is described in creating a data governance committee for IFS Cloud data mesh.
The council changes the culture only if its rules are enforced by the platform, not by reminders. Access to IFS Cloud data is one place where this is already possible: permission sets in IFS Cloud grant access to projections, so an agreed rule about who may read a domain’s data becomes a grant that can be reviewed. The automated side of governance is the subject of apply federated computational governance.
6.How do you scale the change beyond the first domain?
The change scales through a lighthouse domain: one domain with a willing owner and a visible problem, such as spare parts inventory data that maintenance, purchasing and finance all read differently. When that domain publishes a product the others start to use, it becomes the internal example that the next domain leader can check for themselves.
Three habits keep the momentum after that. Early adopters present their product and its consumers at the governance council, so recognition is tied to use and not to effort. Ownership goes into role descriptions and objectives, so it survives a change of manager. And the central team’s success measure changes from reports delivered to domains served by the platform.
If you are planning this change in an IFS Cloud organisation and want an outside view of where to start, see my services for IFS Cloud.
7.Frequently asked questions
What is the cultural shift in data mesh?
The cultural shift in data mesh is the move of data accountability from a central data team to the business domains that produce the data. Domain leaders own their data products and their quality, product owners manage them for consumers, and the central team runs the platform instead of owning the content.
Does data mesh change our IFS Cloud licensing?
Data mesh is an organisational and architectural approach, not an IFS Cloud module or add-on. Any licensing question comes from the tools you add or the users who get new access, and it should be checked with IFS or your partner for those specific changes.
What is the first step of the cultural change for data mesh?
The first step of the cultural change for data mesh is assessing organisational readiness: finding domain leaders who accept accountability for their data, have consumers outside their domain and can allocate team time. This comes before any technical restructuring, because a platform without owners returns to central control.
Why is federated data governance better than central governance in a data mesh?
Central data governance makes one team approve rules for data it does not produce, which slows decisions and misses domain context. Federated data governance lets the domains closest to the data apply quality rules themselves, while a council of domain representatives agrees the few global standards, such as shared identifiers, that make data from different domains combinable.
What should data product owners be trained in?
Data product owners should be trained in product thinking: identifying consumers, writing business definitions, setting and measuring service levels for freshness and accuracy, documenting the product for the data catalog, handling consumer feedback and managing breaking changes with versioning.
8.About the author
Dariusz Myśliwiec: 18 years hands-on with IFS (Apps 7.5–10 and IFS Cloud). IFS Certified Associate Consultant. PRINCE2® 7. Based in Kraków, delivering remotely across Europe and globally through an independent practice. You talk to the consultant who does the work, not a sales layer.
Selected clients from my IFS work: Fugro · LGC · BVI Medical · Betafence (PRÆSIDIAD) · Barlinek · NGK Ceramics · Newag · Oleofarm.
IFS is a registered trademark of IFS AB; this practice is not affiliated with IFS AB.
Which of your domains is ready to own its data?
Bring the list of business areas that use IFS Cloud and the reports that cause the most disputes. On a call I will go through the readiness questions with you and point to the domain that can be your first example.
Book a call about data mesh readiness Read the data mesh overview