Pricing FAQ for the new Microsoft Purview Data Catalog

The new pricing model is substantially different from the old Azure Microsoft Purview pricing model. In the new Microsoft Purview Data Governance, there's no cost for scanning data assets into the data map, but there are charges for 'governed assets' and 'data management.'

When will the new pricing model go into effect?

The new pricing model will go into effect November 1, 2024. The first bill will be received December 1, 2024.

Will I be charged for data governance between September 1, 2024 and November 1, 2024?

Data catalog and data management in the new data governance catalog are free or won't be charged between September 1 and October 31 2024. However, to use data governance within this time period, you'll have to pay for the scan and data map, as you do in the current experience.

Will I be charged for scan and data map from November 1, 2024?

As the new business model for data governance goes live, scan and data map costs will be zeroed out starting November 1, 2024. That said, if you use self-hosted integration runtimes (SHIR) or private endpoints, you'll pay for those meters separately, while scan and data map costs won't be charged.

Is the asset counting done at a monthly level or daily level?

Assets are counted on a daily basis. Each day unique governed assets are counted, so if you have ungoverned any assets mid month, the right asset count will be picked for the day and multiplied by unique governed assets per day.

I have a single data product with three data assets attached, will I be charged for each data asset for 30 days?

Correct, when attaching three data assets to a data product, you'll be charged for each data asset each day.

I have the same data asset attached to both a data product and a critical data element, how will I be charged?

You won't be double charged for a data asset attached to both a data product and a critical data element. Once a data asset is charged, it will not be charged again.

I have scanned 10 data assets and only 3 are attached to data products, how will I be charged?

You'll be charged for the three data assets attached to the data product. You won't be charged for the remaining seven data assets in the data map, as these aren't governed.

If I'm using the classic glossary in the data map and have attached classic glossary terms to a data asset, will I be charged?

Assets with classic glossary terms attached to them won't be charged.

If I attach a SQL server (with 200 tables in it) to a data product, how many governed assets will I have to pay for?

You'll be charged for a single asset as governed asset for a SQL server with many tables. Children won't be counted since the intent is to govern the server as a whole. Although we recommend individual or a group of tables we put in data product, so the experience is more meaningful. If the data product is data consumer facing, then limited data consumers would want access to a data product with 200 tables.

If I attach a Power BI semantic model into a data product that has 200 tables, and then I intentionally hand-pick 10 tables to be part of the data product, then how many assets do I get charged for?

If the semantic model is still attached to the data product, we'll charge it as a single asset. If you have picked five tables to be attached to the data product, removing the semantic model we'll charge for the five tables only.

If only the asset type attached is counted as a governed asset (meaning a Power BI dataset or semantic model is a governed asset), will that restrict me from using the child table for data quality rule set up?

No. Counting assets for billing is separate from user experience and set up of data quality scans. While we'll count the semantic model as a governed asset, when a data quality scan is set up, it will separately accrue billing towards the data management processing unit meter.

Can I run data quality rules on a semantic model or its individual tables and columns?

Data quality runs are set on columns of tables within data products or on a critical data element (CDE) directly.

What type of billing transparency will I be able to see?

There are plans to create a Data Governance Admin view of how much is being billed for data governance usage. For example, N# of jobs using Y# of processing units. Based on the processing units, you can decide if you're over-using, under-using, or if there are opportunities to optimize data quality rules.

What guardrails exist to manage cost overruns?

There will be the Data Governance Admin view, and we'll also provide alerting mechanisms when charges cross a certain threshold. This is a future roadmap item, and not immediately available at November 1, 2024.

Is there a feature to estimate cost for different jobs?

We have a cost estimation feature on our roadmap. The first area is to give you clarity on usage based on the last week of so. This is Azure based billing, so you can see projections in Azure cost to understand whether you're over or under using your resources. The second area is to provide granularity at the job level. We're currently building this along with forecasting future projections.

Are there tools and guidance to help us estimate costs with the new pricing model?

We're working on a directional pricing calculator that will help you with rough estimation of their TCO.

Is there a free trial available to understand spending?

We don’t have a free trial for the product today. However, after we announce GA on September 1, 2024, you won’t be billed for the new data catalog and data management (data quality and data health management) until November 1, 2024 so that period can be used to test and create estimations.

Will there be a volume discount?

Volume based discounts aren't available on November 1, 2024. Based on usage at scale, we can work on volume-based pricing for future releases.

Is there a way to separate costs for charge back purposes at the governance domain level? If so, what exactly is included in these costs?

To rephrase the question, can you separate cost per governed asset at the governance domain level, and data management processing unit charges per governance domain? You can't currently separate cost by governance domain. We're working on a consumption report that will be available to the data governance admin and governance admins. In the report, we intend to provide total governed assets by domain and data governance processing units consumed by governance domain. This should enable you to see volume of use by meter types and use that to charge back within their organization.

What cost reporting will be available s between September 1, 2024 and November 1, 2024, and where will this be accessible in the solution to view?

Between September 1, 2024 and November 1, 2024 there's no cost reporting available. If you need further information, reach out to your CSA and CXE team.

What meters are zeroed out from the classic Microsoft Purview Data Catalog?

  • Data map population
  • Data map enrichment
  • Data map consumption
  • Data estate insights