Independent IFS Cloud practice · Data mesh and master data
Enable a data mesh in IFS Cloud by publishing master data first
A data mesh in IFS Cloud is easiest to start from master data, such as suppliers and parts. Every business domain reads these records, IFS Cloud is the system where they are created and maintained, and their structure changes less often than the structure of orders and transactions, so a master data product gives other domains something stable to build on.
This article answers one question: which master data to publish first, who owns it, and what has to be true before it can be called a data product. The general model of data mesh is in what is data mesh, and the two follow-up steps are data domain mapping and master data as data contracts.
Key takeaways
- Master data is the first data product because every other domain consumes it: a purchase order, a customer order and an invoice all point to a supplier, a customer or a part.
- One master data record in IFS Cloud can have several owners. A supplier carries purchase defaults (payment terms and currency, among others) and invoice defaults (currency, payment terms, tax code).
- Site is part of the key of an inventory part in IFS Cloud: the 24R2 analysis model documentation states that “each contract (site) has its own set of inventory parts” (IFS Cloud 24R2). A part entered with no site is not tied to any site. The product owner and the inventory owner answer for different records.
- A master data record becomes a product only when it has a named owner and written definitions. It also needs measurable quality rules and an agreed interface. Publishing a projection without them publishes a table.
- The sequence is: choose the master data, map it to owners field group by field group, then write a contract for each product.
1.Why should a data mesh in IFS Cloud start with master data?
A data mesh should start with master data because master data is the part of the IFS Cloud data estate with the most consumers and the clearest system of record. Transactional data mostly serves one process. A purchase order matters to procurement and receiving, and to accounts payable. A supplier record matters to all three of them as well as to quality control, to finance reporting and to any analytics model that groups spend by supplier.
Three properties make master data the right first product:
- It is created in one place. Supplier and part records are entered and validated in IFS Cloud, so there is no argument about which system holds the authoritative version.
- Its structure is stable. New fields appear, but the record itself rarely changes shape, so an interface built on it does not break every month.
- Errors spread. A wrong payment term or a duplicate supplier reaches every order and invoice that uses it. Fixing the master record fixes the downstream data in one move.
The reverse order fails in a predictable way. A team publishes order data as its first product, consumers join it to supplier and part data they took from somewhere else, and the first quality dispute turns out to be about master data nobody owns.
2.Which master data domains come first, and who owns them?
The first master data domains are the ones behind the decisions the data mesh vision names. When the vision does not settle the order, I recommend starting with supplier, because every purchase order and every supplier invoice points to it, and its shared fields, payment terms and currency, force the first ownership decision between procurement and finance early. Product follows, then customer and asset. The table below is illustrative for a manufacturing company on IFS Cloud, not a client record.
| Master data domain | Typical owner role | Main consumers | What to settle first |
|---|---|---|---|
| Supplier | Purchasing manager | Accounts payable, quality, spend analytics | Who owns payment terms and currency, which exist both in the purchase defaults and in the invoice defaults |
| Customer | Sales manager | Accounts receivable, logistics, service | Who may create a customer, and who answers for credit and invoicing details |
| Product (parts) | Engineering or product management | Planning, purchasing, sales, inventory | Which data belongs to the part with no site, and which to the inventory part at each site |
| Asset | Maintenance manager | Finance, operations, service | One definition of what counts as an asset record, and who retires it |
Two facts from the IFS Cloud documentation explain the “settle first” column. For suppliers, the general supplier purchase information holds defaults for purchase orders, for example “a supplier statistics group, a buyer, any additions, payment terms, and currency” (IFS Cloud 24R2, Enter General Supplier Purchase Info), while the supplier invoice information holds the currency, the default payment terms and the tax code (IFS Cloud 24R2, Enter Supplier Invoice Information). For parts, the Enter Part Data activity creates a part with no site that is later registered as an inventory, purchase or sales part (IFS Cloud 25R1, Enter Part Data). Both are places where one business object has more than one natural owner.
How to identify the domains themselves, and why company and site are keys inside a domain rather than domains of their own, is covered in identify data domains.
3.What must be true before master data counts as a data product?
Master data counts as a data product when a consumer in another domain can use it without asking anyone what a field means or whether a record can be trusted. In practice that takes five conditions.
- A named owner. One person per domain answers for the quality of the data, for access to it and for changes, and the sponsor has confirmed the role.
- One record per real-world object. Duplicates are found and merged, and a rule says how new records are checked. IFS Cloud has a configurable duplicate check, but its documentation scopes it to leads, accounts and supplier requests (IFS Cloud 25R1, Duplicate Check), so for other master data the rule has to be written and checked by the domain.
- Written definitions. Status values and classification codes mean one thing, and so do units, documented where consumers find them. The golden record article covers how to decide which version of a record wins.
- Quality rules with a measure. For example: every active supplier has payment terms in both the purchase and the invoice defaults, and they match.
- An agreed interface. Consumers read the product through a named interface with a version, not through a database extract somebody built once.
Each condition is organisational before it is technical. No IFS Cloud setting assigns an owner or decides what “active supplier” means.
4.How does IFS Cloud serve a master data product to other domains?
IFS Cloud serves data to other domains in two ways: through projections for operational use and through the Data Services stack for analytics.
| Channel | What it is | Suits |
|---|---|---|
| Projections | The OData services used by IFS Cloud Web, which the API Explorer documents with OpenAPI specifications | Integrations and applications that need current records |
| Data Services | Per the 24R2 documentation, “The Data Pump will do the actual data movement and Parquet file generation”, and the main data store is a Data Lake (ADLS Gen 2) | Analytics and reporting that read history in bulk |
The Data Services description is in the IFS Cloud 24R2 Data Services documentation. The projection side, including how to find which projection a page uses, is described in section 5 of what is data mesh. The channel is a delivery choice for the product. The ownership and the definitions stay the same whichever channel a consumer uses, and so do the quality rules.
5.What is the order of work for the first master data product?
The first master data product takes three pieces of work, in this order, and each has its own article in this series.
- Choose the domain. Pick one master data domain with a willing owner and a consumer who is waiting for it, using sections 2 and 3 above.
- Map it to owners. Assign every field group of the record to one owning domain and record who reads it. The method is in data domain mapping.
- Write the contract. Put the interface and schema into a machine-readable contract. Add quality rules and access rules, then the versioning policy. The method is in IFS Cloud master data as data contracts.
The agreement between the producing and the consuming domain, which the contract then formalises, is covered in confirm sharing agreements.
6.Where does a master-data-first data mesh go wrong?
A master-data-first approach goes wrong when master data is treated as an IT asset instead of a business one. These are the patterns I check for:
- One owner for all master data. A central master data team owns every master data domain at once. It becomes the queue that data mesh was meant to remove, and the business teams that create the records stop feeling responsible for them.
- Cleaning outside IFS Cloud. Duplicates and wrong values are fixed in a spreadsheet or in the data lake, while the source records in IFS Cloud stay wrong and the next extract brings the errors back.
- Shared fields left unassigned. Payment terms on a supplier are edited by purchasing and by accounts payable, and when they differ nobody knows which value is right.
- Everything at once. All master data domains are published together before the first one has an owner, definitions and quality rules. Start with one.
A migration to IFS Cloud is a natural moment to settle ownership, because every master record is reviewed and someone has to approve each cleansing decision. That angle is covered in data migration as a business reset.
7.Frequently asked questions
Why start a data mesh in IFS Cloud with master data instead of transactions?
A data mesh in IFS Cloud starts best with master data because master records such as suppliers and parts are read by every domain. They are created and validated in IFS Cloud, and their structure changes rarely. Transactional data such as purchase orders depends on master data, so publishing transactions first leaves consumers joining them to master data that nobody owns.
Is data mesh only for large enterprises running IFS Cloud?
Data mesh is not limited to large IFS Cloud customers. A mid-sized company can apply the same rules to one or two master data domains. Each needs a named owner and written definitions, and each needs measurable quality rules and an agreed interface. The scale of the platform differs, but the ownership rules are the same.
What is the biggest obstacle to a master data product in IFS Cloud?
The biggest obstacle to a master data product in IFS Cloud is ownership. Moving from “IT owns the data” to “the purchasing manager owns supplier data” needs a decision by the sponsor, time allocated to the owner, and a rule for shared fields such as supplier payment terms. IFS Cloud provides interfaces and access control, but not the decision about who answers for the data.
Does IFS Cloud have a master data management module for data mesh?
IFS Cloud does not ship data mesh as a module. Master data such as suppliers and parts is maintained in the standard IFS Cloud functional areas and served through standard mechanisms: projections and permission sets, plus the Data Services stack. The organisation decides ownership and definitions. It also sets the quality rules.
How does a supplier record in IFS Cloud end up with two owners?
A supplier record in IFS Cloud holds purchase defaults, such as the buyer and the payment terms used on purchase orders, and invoice defaults, such as the default payment terms and the tax code. Purchasing works with the first group and accounts payable with the second. The domain map has to name one owner for each field group, and one owner for fields that appear in both, such as payment terms.
8.About the author
Dariusz Myśliwiec: 18 years hands-on with IFS (Apps 7.5–10 and IFS Cloud). 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.
If you want help choosing and preparing the first master data product in your IFS Cloud, see my services for IFS Cloud.
IFS is a registered trademark of IFS AB; this practice is not affiliated with IFS AB.
Name the owner of your supplier payment terms
If purchasing and accounts payable both edit payment terms on your suppliers, the first master data product in your IFS Cloud already has its first open decision. A short call is enough to pick the domain and the owner.
Book a call about your first master data product Read the next step: data domain mapping
