“As D&A professionals we should always be asking ourselves not just whether we are collecting, integrating and storing our data in the best way possible, but more fundamentally, whether we have the right data.”

“Having the right varieties of data is more important than volume.”

"D&A leaders can ensure they have the right data by doing four things: 1) stop collecting data just in case; 2) consider substituting big data for small data, 3) swap real data for synthetic data, and 4) ensure your active metadata tells you not only what data you have, but also what it means."

“The most valuable data will be the data we create, not the data we collect.” 

"Gartner estimates that by 2030, the majority of the data used to build models will be synthetic data."

“Data alone is unlikely to drive decisions. Design better decisions by improving the timing of decisions, accelerating decisions and connecting decisions.”

Learn more in the associated Gartner press release.

“Have your long-term vision: enabling outcome-focused DSML. In the near-term, be intentional: aim for an MVP, skip a POC. This will get you closer to the valuable outcomes.”

“Do not overcomplicate solutions to deliver value. Use the best technique for a job to achieve efficiency, speed and simplicity. Leap to the state-of-the-art ML with pretrained models when it gives you a competitive differentiation.”

“Manage risk, support accountability and increase adoption with Responsible AI tools.”

“Optimize for bottleneck skills: enable professional data scientists to do the highest value work.” 

“Attract talent with new career opportunities: involve engineers in designing AI solutions, operational efficiencies and implementation strategies.”

“Build a data foundation for AI. Approach data-centric DSML as an ongoing, high-priority investment.”

“Once a data pipeline is up and running, it can fail due to changes to data schemas, new values being introduced or anomalies suddenly appearing in data values. These actions result in bad data being delivered to downstream processes.”

“DataOps is emerging as a response to similar friction around the consumption and use of data across the organization.”

“There are three best practices D&A leaders must follow to improve data engineering success: 1) Replace monolithic practices with modularity, 2) Pick the right use case for automation using metadata analytics, and 3) Implement an enablement mindset by deploying agile practices for citizen roles.”

“Create data engineers by upskilling your ETL developers, data analysts, DBAs or similar roles.”

“Create mixed-role teams, train members if you must.”

“Embrace automation wherever possible to accelerate design, development, monitoring, and management of data products that meet your business demands.”