Key Takeaways
Here are key takeaways from the event contributed by Kevin Petrie and Wayne Eckerson in the final session.
Keynote Takeaways:
- Data products are more than data assets
- They are reusable, shared, and continuously funded, with target users and rigorous data governance. Can be code or data
- Really needs a data market – like a catalog of data products – unlocks value with frictionless data sharing
- Data product platform: product studio, storage, sits on top of fabric
- Build, buy, mix, match
- Many data products will be virtual
- Data commerce, integration, catalogs, analytics, MDM vendors all moving into the data product platform space as well
- Data product manager is product manager, owner, and marketer
- Data product organization can be centralized or decentralized – pros/cons to each
- Data product governance requires careful organizational structures and workflows, with review/approval process, oversight, standards, etc.
Practitioner Takeaways:
Mike Ross, SVP of Technology, Blue Sky Specialty Pharmacy
- Data products must be testable, usable, and loveable
- Product managers have product management experience. Paired with product counsel, which has senior leaders of company. They drive conversations with outside users
- Journey from query to Excel file to application that is continuously used
- They’ve matured to go through the full lifecycle – even have retired a data product
Anupam Nandwana, CEO P360
- Data products can derive from MDM golden records
- Key barriers: data volumes, product ownership, cross-functional processes for standardizing
Featured Presentation Takeaways:
Rick Hall
- More than half of data leaders cite issues with data quality, reusability. This contributes to huge turnover among CDAOs
- Four principles:
- Delivery of quality granular data products that might otherwise be overlooked
- Empowered business teams must move quickly
- Curate results to identify trusted content
- Foster collaboration and reuse
Anthony Deighton
- data products are consumption-ready, high-quality, trustworthy, accessible, and able to help solve business problems
- Clean, standardized, many attributes, with consistent views but also the ability to iterate
- Data teams can’t keep up
- Historical solutions are to throw people at the problem or lock it down – i.e., MDM or other controls prevent future data changes
- AI can unlock and break the dilemma
- Start with a MVDP
- Nobody likes being on an MDM team…
Expert Panel:
- Data governance cannot be purely centralized/top down (Rick)
- Start with consumption rather than production (Anthony)
- Data mesh philosophy reminds us that data product ownership should be as close to the source as possible (Rick)
- You need both central and distributed elements – i.e., a federated model
- Empower but govern edge data consumption and usage
- Value metric focuses downstream – i.e., the output and what consumer does with it
- Costs are processing, time, etc.