A data product can have every field filled in and still go unused. Completeness is not the same as usability: what decides whether other teams adopt a product is whether they can find it, understand it, and trust it without talking to you first.
Six characteristics separate a product people adopt from one that merely exists. They come from the original Data Mesh definition of a data product and are often shortened to DATSIS. Each one maps to something concrete in Dawiso, so “good” is a checkable property rather than an opinion.
| Characteristic | The question it answers | What makes it true in Dawiso |
|---|---|---|
| Discoverable | Can someone find this without knowing it exists? | Published to the Data Market, with tags, categories, and a real description |
| Addressable | Can they reach it at a stable location? | A stable product ID and an output port with an access URL |
| Trustworthy | Can they rely on it? | A named Owner, a production workflow state, quality checks, and an SLA |
| Self-describing | Can they understand it without asking you? | A data contract with schema, plus a value proposition and use cases |
| Interoperable | Does it work with other products? | Typed ports, shared classifications, and export to open standards |
| Secure | Is access controlled and appropriate? | Contract access roles, visibility, and sensitivity classification |
Discoverable
A product nobody can find has no consumers. Discoverability means someone who does not know your product exists can still come across it while looking for the data they need.
In Dawiso, a product becomes discoverable when it reaches production and appears in the Data Market. Tags and categories place it alongside related products, and the description and value proposition are what a stranger reads first. A product tagged with nothing and described in one line is technically published but effectively hidden.
Addressable
Consumers need a stable way to reach the data — one that survives internal reorganizations. Addressability is the difference between “ask the Marketing team where the churn table lives” and pointing at a fixed location.
Two things provide this: the product ID, which never changes once set, and the access URL on each output port. Because consumers depend on the port rather than the assets behind it, you can move or rebuild the underlying tables without breaking the address.
Trustworthy
Trust is what makes someone build on your product instead of copying the raw data. It comes from evidence, not assurance.
Four things supply that evidence: a named Owner who answers for the product, a workflow state of production that signals it is safe to depend on, quality checks that show the data is tested, and an SLA that states what freshness and availability to expect. A product with none of these asks consumers to take a risk they cannot size.
Self-describing
A consumer should be able to evaluate and use your product without a meeting. Self-describing means the product carries its own documentation.
The data contract does most of this work: schema objects and properties tell a consumer which fields exist, what they are called, and how they are typed, before a single row is read. The value proposition explains what the product is for, and use cases show what it has been used for. Field descriptions matter more than they look — an unexplained column is a support request waiting to happen.
Interoperable
Products become more valuable when they combine. Interoperability means yours can be joined to others without custom glue.
Typed ports let one product consume another directly, which is what makes cross-product lineage possible. Shared classification code lists mean sensitivity means the same thing everywhere. Because the model follows open standards, a product also stays readable outside Dawiso.
Secure
Access has to be deliberate. Secure means the right people can use the product and the wrong people cannot, and that both are visible.
Access is granted through the contract’s roles rather than by sharing credentials, so every grant has an approver. PII Level and Confidentiality tell consumers and reviewers how the data must be handled, and visibility controls who sees the product at all. Classification also protects you: an unclassified product carrying personal data is a compliance problem nobody has noticed yet.
Judge your own product
Read the six characteristics as questions about a product you have already built. The ones that are hardest to answer are usually the ones consumers would have struggled with too.
For the specific items to verify before publishing, see Pre-publish checklist. For the failure modes that break these characteristics in practice, see Data product anti-patterns.