As businesses increasingly depend on data to drive innovation and growth, poorly executed—or completely absent—data product strategies have emerged as a major strategic problem.
Imagine this: You’re tasked with moving valuable freight across the country. Would you attach a separate engine to each cargo car? Of course not—you’d link multiple cars to a single engine for efficiency and scalability. You’d also standardize trains and connectors to move various goods easily.
The same logic applies to data products. Real scale and value come from treating data products as engines that can support multiple use cases—not as isolated, one-off efforts. Yet many companies still operate under a “one engine, one car” mindset, resulting in fragmented, low-value data initiatives that stall progress.
When we first explored this in 2022, we outlined how managing data like a product can unlock broad benefits. Since then, adoption has grown, spurred even further by the excitement around generative AI (gen AI). Executive teams increasingly recognize the need to better harness their data assets.
However, the results have been mixed. Confusion about how data products create value, governance focused on individual projects rather than strategic gains, and incentive structures that reward building over scaling have all limited success. As companies lean more heavily on data for initiatives like gen AI and digital twins, the ability to properly develop and scale data products is becoming business-critical.
Through working with dozens of organizations, we’ve learned that creating valuable data products is less about technology and more about leadership and operational discipline. Five key lessons emerge:
1. Focus on Value Creation, Not Just Better Data
The point of data products isn’t to create “better” data—it’s to drive business value. No data product initiative should begin without a clear understanding of the business value each use case offers. Leaders must prioritize opportunities by assessing the projected returns over the next 12 to 24 months and grouping similar needs together.
If a use case stands alone, building a full data product may not make sense. But when multiple valuable use cases share underlying data, it’s worth the investment. Mapping use cases alongside their expected returns provides a strategic guidepost for leadership alignment and decision-making.
Often, realizing value means building multiple complementary data products. For example, a telecom company optimizing its network had to build separate products for technician data and tower performance to complete the necessary analysis.
2. Understand the Economics Behind Data Products
Many organizations lack a solid grasp of how data products generate returns. A data product’s true value lies in its ability to accelerate value capture and lower marginal costs with each new use case.
Initial development is expensive—often 60–80% of the effort is a one-time setup—but reusing the data product for future projects amortizes these costs over time. For instance, an international consumer brand cut costs by 30–40% by scaling an existing data product to multiple use cases and markets, instead of creating new data pipelines each time.
Beyond cost savings, well-built data products speed up realization of business value—sometimes by as much as 90%. They also significantly reduce risks and defects by improving data quality and standardization.
Key takeaways:
- Most value often comes from just five to fifteen core data products.
- CIOs must clearly communicate data product ROI to justify investments.
- Teams should focus less on churning out new products and more on growing and maintaining valuable ones, with incentives tied to meaningful metrics like reuse rate, user satisfaction, and maintenance efficiency.
3. Build Data Products Designed to Scale
Strong engineering underpins scalable data products. Too often, companies underinvest in this foundational work, leading to costly failures.
Essential principles for scalable data product design include:
- Future-proofing: Anticipate future data needs. Build products flexible enough to incorporate new sources without major rework.
- System integration: Use standardized APIs and connectors so products integrate seamlessly into existing IT systems.
- Ease of access: Create a user-friendly marketplace (similar to an App Store) where internal users can easily find, access, and request support for data products.
- Automation: Invest in DataOps to automate tasks like data rationalization, lineage tracking, and security. Automations must be reviewed periodically as business needs evolve.
- Reliable data sourcing: Create curated, standardized domains (like supply chain data or customer data) to ensure reliable inputs.
Skilled data engineers are critical. Great ones are not just technical but also business-savvy—asking deep questions and solving problems creatively.
4. Appoint Business-Savvy Data Product Owners (DPOs)
Viewing data products as mere IT projects misses the mark. They must be treated like business assets managed by capable leaders.
- DPOs should run the show. Too often, project managers are assigned to lead, causing a focus on deliverables rather than value creation. Strong DPOs actively identify new use cases, track KPIs, drive adoption, and manage costs. Some organizations even tie DPO compensation to the business value they help create.
- Separate domain and product management. Data stewards (not DPOs) should oversee domain integrity and quality standards.
- Engage the business early. Subject matter experts and operational leaders must be involved from the start to ensure that data products address real business needs.
At the organizational level, success often hinges on senior business leaders—not data teams—championing efforts to align people, processes, and priorities across departments.
5. Leverage Generative AI in Data Product Development
Gen AI can accelerate data product development up to three times faster than traditional approaches. However, many companies fail to harness its full potential.
Areas where gen AI can add value include:
- Drafting user stories and acceptance criteria
- Generating requirements based on business needs
- Building data relationships and transformation code
- Automating testing for quality and privacy
Additionally, gen AI opens access to new types of unstructured data (like images or reviews), expanding data product capabilities. But to capitalize on this, companies must organize and tag their unstructured data intelligently to keep access efficient and costs manageable.
Conclusion
In short, raw data is like crude oil—useless without refining. Data products are the refineries that unlock business value. But merely building them isn’t enough. Companies must scale them systematically, with a clear focus on strategic value, sound economics, solid engineering, strong leadership, and the smart use of gen AI.
Only then can organizations fully realize the promise of data-driven decision-making and innovation.
The views and opinions expressed in this article are solely those of the authors and do not necessarily reflect those of Bespoke Business Development. They are intended to encourage discussion and reflection, rather than serve as legal, financial, accounting, tax, or professional advice.
Have Questions or Thoughts About Our Latest Insights?
We’d love to hear from you. Whether you’re curious about a recent post, want to explore a topic further, or have ideas you’d like to share—reach out to us. Our team is here to connect, collaborate, and provide clarity.
