One unit, one owner
Data, transformations, quality rules, contract and access pattern travel together under a named owner. Not a table someone left behind in a schema.
Qlik Product · Data products and Trust Score
Most data lands somewhere and then belongs to nobody. A data product bundles the data, the transformations, the quality rules and the ownership into one unit, and carries a score that tells you how far to trust it before you build on top of it.
A data product in Qlik Talend Cloud is a governed unit: the datasets, the transformations that shaped them, the quality rules they have to satisfy, the access pattern, and a named owner. It is published to a marketplace where the rest of the organisation can search it, read its documentation, see where it came from and see how good it is, then use it in Qlik Cloud Analytics without asking anyone for a copy. The Qlik Trust Score is the number attached to that last judgement. It rates a dataset on five axes and updates whenever something happens that changes one of them. Validity and completeness are computed on the data itself and cannot be switched off; the weighting across the five is yours to set.
Core capabilities
Data, transformations, quality rules, contract and access pattern travel together under a named owner. Not a table someone left behind in a schema.
The Qlik Trust Score rates validity, completeness, popularity, discoverability and usage, weighted the way you decide, and updates itself as things change.
Consumers find a product by keyword, read its documentation, check its lineage and profiling, and see the score before they build anything on it.
Trace a figure back through its transformations to its source, and see what breaks downstream before you change it rather than after.
The Trust Score
Five axes, each answering a different question about a dataset. Validity and completeness always count and cannot be turned off. The weight of each is yours to set, and the total comes to 100%.
| Axis | What it looks at | Why it matters |
|---|---|---|
| Validity | The share of values matching the expected type and semantic format | A column full of almost-right values is the most expensive kind of wrong |
| Completeness | The share of records that are actually filled in | Gaps decide whether an average means anything at all |
| Popularity | User ratings and the certification level the dataset carries | What the colleagues who use it every day already know about it |
| Discoverability | How well it is documented: descriptions, tags, custom attributes | An excellent dataset nobody can find gets rebuilt by somebody else |
| Usage | How often it feeds pipelines and preparations, and how often it is opened | Something the business runs on carries a different risk from something nobody touches |
Quick quote
Data products and the Trust Score are part of Qlik Talend Cloud. Fill in the form and we will work out which edition and capacity fit, including a first domain to pilot on.
Frequently asked
No. A dataset is a table. A data product is a table somebody is answerable for: it has a named owner, documented content, quality rules it has to keep meeting, an access pattern, and a lifecycle that continues after the project that created it ends. That accountability is the whole point, and it is also the part no tool can install for you.
Yes. You set the weight of each of the five axes and the percentages total 100%. Validity and completeness cannot be switched off, because they are the two that measure the data itself rather than how people treat it. In practice we start from the Qlik default and adjust after the first review with the domain owners, once you can see what the scores are actually telling you.
Yes. Data products and the Trust Score live in Qlik Talend Cloud, which is where the integration, quality and governance work happens. Consumption is separate: once a product is published, people use it from Qlik Cloud Analytics like any other source, with the score and lineage visible next to it.
The tooling is the easy half. The hard half is deciding which domains own which products, what good enough means for each one, who is called when a score drops, and how that is enforced without turning into a committee. That is DAMA DMBoK work and it is what we do day to day. We usually start with one domain, get a working product published, and let the rest follow that pattern.