Master data creates real technical and organisational challenges. Choosing the right data hub is how you solve them — here are the four models and how to pick one.
Building a single master data repository is never a trivial exercise. It raises significant technical and organisational challenges, and the most common of them are familiar to almost every data team:
- Defining a single data model is hard. Business data models depend on how each department actually uses the data, so a definition that satisfies everyone rarely comes for free.
- Mapping data properly takes time. Establishing the central model — its attributes, structure, and relationships — is slow, detailed work.
- MDM needs ongoing supervision. Sound data governance has to run continuously, and that too costs time and effort.
- MDM reshapes the organisation. Teams often have to review — and change — the way they work with data.
Because organisations differ so widely, Master Data Management (MDM) takes different shapes, and several architectural models exist. Depending on how a model is built, it can enforce regulatory requirements, grant local sites real autonomy, or centralise business-critical data under tight control. In other words, MDM can adapt to the way your teams already work and to the technical and regulatory constraints of your industry — rather than the other way around.
What Is a Data Hub? The Core of Any MDM Architecture
At the heart of every MDM architecture sits the data hub: the component that aggregates data and makes the single master data repository easy to query. How the hub handles that data — whether it merely indexes it, physically consolidates it, or takes ownership of it — is what distinguishes one MDM model from the next. There are four main data hub models.
Build a trustworthy data foundation to enable AI-driven automation and reliable decision-making. Download the whitepaper to learn how to establish Data Governance as a strategic capability — and unlock scalable automation and AI.
1. The Registry Data Hub
The registry model is comparatively simple to implement and leaves data in the ownership of the source systems. The hub takes responsibility for de-duplication and data cleansing, then cross-references those results to establish a single version of the truth. Continuous cross-referencing produces clean data indexed by unique keys. That data is not written back to the source systems; it is available read-only.Advantages
- A single model is easy to build from multiple source systems
- Nothing is deleted from source systems, which sidesteps compliance and regulatory risk
- It is non-intrusive and quick to roll out
Disadvantages
- It does not produce a physically consolidated data set
- It is relatively inefficient at large data volumes
2. The Consolidation Data Hub
In the consolidation model, data is copied from the source systems and consolidated into a golden record inside the hub. That golden record can then be redistributed to applications or consumed directly by business units. Because it provides trustworthy master data, reporting becomes far easier. Unlike the registry model, the consolidation hub also feeds data back to source systems to enrich and update what they hold — so applications and processes can keep improving the quality of their data.Advantages
- Genuine consolidation inside the hub, enabling a comprehensive master data repository
- Economical and reliable for all kinds of reporting and analysis
- Less dependent on source systems than the registry model, since data is centralised in the hub
Disadvantages
- Long intervals between consolidation cycles can leave data in the hub out of date
3. The Coexistence Data Hub
The coexistence model builds directly on consolidation: it creates a golden record, then redistributes the cleansed data back to applications, where it is integrated. Source systems keep control of their data, but the constant updating usually pushes business departments to adapt their methods. Frequently used as a transition architecture when moving from a registry hub toward a fully centralised model, coexistence spreads more unique, reliable data across the entire application landscape.Advantages
- More trustworthy, unique data in both the hub and the source systems
- Faster access to data, which improves process performance and simplifies reporting — well-defined attributes appear in reports immediately
Disadvantages
- More intrusive than the registry and consolidation approaches
- Integration carries more technical and financial constraints: data models must be well designed and clearly structured before MDM can use them
4. The Transactional (Centralised) Data Hub
The centralised model places master data firmly under the authority of the MDM hub. The hub becomes the data supplier and the single reference point for every source system. Continuously enriching, de-duplicating, and cross-referencing, it maintains up-to-date master data and dispatches it out to the connected systems.Advantages
- Truly unique master data that can be trusted at any moment
- Strong security and compliance across data processing
- Systems and processes benefit from direct data enrichment
Disadvantages
- Highly intrusive: existing processes usually have to be redesigned
- Integration is more complex, more costly, and slower to deliver
Comparing the Four Data Hub Architectures
Which model fits depends on how much control you want the hub to hold, how fresh the data must be, and how much change your organisation can absorb.| Model | Data ownership | Writes back to sources? | Intrusiveness | Best suited to |
|---|---|---|---|---|
| Registry | Source systems | No (read-only) | Low | Fast, non-intrusive de-duplication across many sources |
| Consolidation | Shared (golden record in hub) | Yes | Medium | Reliable reporting and analytics |
| Coexistence | Shared, synchronised | Yes | Medium–High | Transition toward a centralised model; unique data everywhere |
| Centralised (transactional) | MDM hub | Yes (authoritative) | High | Single source of truth with strict security and compliance |
SoftProject’s MDM Expertise
Choosing the right data hub is rarely a one-product decision — which is exactly why SoftProject brings together a set of dovetailing capabilities that let you evolve your architecture and support your data strategy on your own terms. Effective data traffic management, intelligent process modelling, and master data management that unifies your data combine into one coherent answer to the business need. On the SoftProject platform, that translates into:- Master data management and data governance — SoftProject MDM unifies your master data, while dataspot. adds business-oriented data governance and data lineage so every golden record traces back to clearly defined, well-understood data.
- A transparent data foundation — MyDataCatalogue catalogues the systems and sources that feed your hub, making it clear where each attribute originates — essential for audits and compliance.
- Data traffic and integration — the X4 BPMS low-code platform connects your applications through an integrated ESB and 200+ ready-made adapters (seamless integration), so data moves cleanly between systems and the hub.
- Process modelling and automation — model, simulate, and automate the surrounding processes graphically with BPMN 2.0, then keep them running and governed at enterprise scale with Phoenix for enterprise integration and execution governance.
Ready to Unify Your Master Data?
The right data hub turns scattered, duplicated records into a trusted foundation for reporting, compliance, and day-to-day operations. Whichever model fits your organisation, SoftProject can help you design it, implement it, and supervise it over time.
Edouard Cante is responsible for the strategic direction and further development of SoftProject’s product portfolio as Chief Product Officer. With a strong understanding of the market and a high level of innovative drive, he advances customer-centric solutions and ensures the company’s long-term competitiveness.
FAQs Data Hubs
What is a data hub in MDM?<
A data hub is the central component of a master data management architecture. It aggregates data from source systems and makes the single master data repository easy to query. Depending on the model, it may index data in place, consolidate it into a golden record, or take full ownership of it as the authoritative source..
What is a golden record?
A golden record is a single, trusted, de-duplicated version of a master data entity — a customer, product, or supplier, for example — assembled from multiple source systems. It gives reporting and operational processes one reliable reference to work from.
Which data hub model should I choose?
It depends on how much authority the hub should hold and how much change your organisation can absorb. A registry hub is fast and non-intrusive but read-only; consolidation suits reliable reporting; coexistence is a common stepping stone toward centralisation; and a centralised hub delivers a true single source of truth at the cost of higher intrusiveness. Many organisations start light and migrate toward centralisation as governance matures.
How does master data management relate to data governance?
MDM is the discipline of creating and maintaining trusted master data; data governance provides the policies, ownership, and oversight that keep it trustworthy over time. In practice the two work together — which is why SoftProject pairs its MDM capabilities with dataspot. for governance and lineage and MyDataCatalogue for cataloguing.
