Trusted Data: Alchemy For Misinformation

neub9
By neub9
4 Min Read

The best description of untrusted data I’ve ever heard is, “We all attend the QBR – Sales, Marketing, Finance – and present quarterly results, except the Sales reports and numbers don’t match Marketing numbers and neither match Finance reports. We argue about where the numbers came from, then after 45 minutes of digging for common ground, we chuck our shovels and abandon the call in disgust.”

How would you go about fixing that situation? How would you get the trust into trusted data?

Consult the Book of Spells

Our spells are cast from our Enterprise Business Glossary. Our wizard is Data Governance Director Suvayu Bose (no relation) who employs a very practical approach to data governance: establish C-suite commitment to the program, set strategic goals, identify data owners and data stewards, then get right to negotiating data definitions cross-functionally.

For data to be trusted, everyone must first agree to what it means, where it’s sourced, and how it’s derived.

Start with critical data elements, those data objects comprising the most important metrics and KPI to run the company.
In this respect, Suvayu is quite the Svengali (no relation). If your numbers don’t conform to his data definitions, you’re up the QBR without a shovel.

  1. Standardize Datasets

Here’s the first of three things Suvayu recommends to get the trust in
trusted data: as data definitions are codified in the business glossary, establish those data objects in your enterprise datasets and evangelize them as the source of truth from which new data assets should be sourced.

Our company built the world’s best hybrid cloud data platform, bundled with integrated security, governance, and lineage, and yet we face the same challenges governing internal data that you might. We doubled-down on data governance in 2021, and in 18 short months we’re flying high, in part because we are standardizing our enterprise datasets. By sourcing new analytics from standard datasets, archiving legacy datasets, and repiping established analytics (only when feasible and purposeful!), we increase trust in data.

  1. Standardize Reporting & Analytics

We’ve been great at data democratization for years but we’ve experienced the common adverse side effects that perhaps you face as well: the ungoverned proliferation of contrary reporting and analytics. Inventory shrinkage increases trust in the data by removing access to duplicative, contradictory reports.

First we retired reports and extract jobs with no/low usage: 85% of the inventory! That exposed additional db archival targets. We built enterprise standard dashboards for the company’s most important KPI and metrics, beginning with executive views then drilling down into middle management and individual contributor views. Then we consolidated an additional 5% of inventory by grafting important features of well-used reports into the enterprise standards.

  1. Standardize Everything In-Between

With enterprise standard data objects and dashboards on the rise and legacy data assets in decline, we shutoff duplicative pipelines and queries and we watched the health of our environment skyrocket.

If you need help (we did), engage our Professional Services team to identify where your opportunities are and how to realize them.

Share This Article
Leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *