DataIQ clients can find the full set of learnings here.
Click here to become a DataIQ client and access all the learnings, plus regular masterclass sessions and other peer-led discussions.
Synopsis
A large global bank is adopting a data product approach. After substantial investments in data platforms and foundational capabilities, expectations are now shifting towards demonstrable business value and faster delivery and adoption. Framing data products as a tiered ecosystem – from raw, experimental data assets through to foundational, business and enterprise-level products, this DataIQ client sought peer advice on balancing speed, democratisation and trust.
Peers from mature, complex organisations shared practical lessons on accelerating adoption, embedding data products into workflows and decisions, structuring teams and funding models, and using speed, rather than perfection, as a strategic lever.
Client Questions
- What does it take to build data products that are actually used by the business?
- How are teams integrating data products into real workflows, decisions, and outcomes?
- What have we learned about the roles, structures, and habits that make these efforts succeed?
Anchor data products to urgent business demand
Peers consistently stressed that adoption improves when data products are built around real business demand, not enterprise ideals. Products gained traction when they unlocked something the business urgently needed, such as group-level visibility during an acquisition or insight that could not be delivered through existing reports.
Attempts to replace legacy pipelines that “still worked” met strong resistance. By contrast, new use cases created pull rather than push. Adoption accelerates when data products remove an immediate blocker, not when they are positioned as cleaner alternatives to familiar outputs.
Prove value early by keeping products deliberately small
Several leaders reflected that early data products failed because they were too broad and abstract to gain traction. Momentum improved when teams aggressively narrowed scope to a small number of high-impact products and worked closely with users. Right-sized products moved faster through modelling and governance, enabled earlier feedback, and made ownership clearer—shortening time-to-first-use while foundational work continued in parallel.
Target business-embedded analysts as the primary adoption channel
Across sectors, leaders were clear that the fastest route to adoption runs through analysts embedded in business functions, rather than central data teams or end users. These insight and analytics teams sit close to decisions, understand the context of the data, and act as an intermediary layer, using trusted data products to create dashboards, analyses and insights for wider audiences.
End users rarely consume data products directly. Instead, business-embedded analysts value consistency, speed and semantic clarity, and are able to work with transparent quality constraints without waiting for perfection. By enabling these teams first through curated data products and governed semantic layers, organisations accelerate downstream value and data democratisation without over-engineering products for every user group.
Integrate data products into workflows via semantic and consumption layers
Many organisations deliberately limit direct access to curated data products, instead embedding them into everyday workflows through semantic models and BI layers. This approach allows analytical teams to self-serve from trusted data products while ensuring outputs are delivered through tools the business already uses, such as dashboards and reports, reducing duplication and protecting trust.
In practice, curated products are consumed via semantic models (such as those from Power BI), while direct access to the lakehouse is tightly controlled through persona-based entitlements and fine-grained row- and column-level policies enforced using tools. While this model accelerates adoption and supports data democratisation where it matters most, peers acknowledged ongoing tension from teams seeking deeper access to raw data, making clear personas, boundaries and governance essential to sustaining trust at scale.
Separate governance, ownership and delivery to remove bottlenecks
Peers agreed that architect-led approaches help establish consistency and trust in complex, regulated environments, but stall quickly when architecture becomes detached from business ownership and delivery. One leader shared a federated operating model designed to avoid this by clearly separating governance, business accountability and information management.
In this model, data governance provides guardrails rather than control, with data officers embedded in business functions, supported by data custodians, and data governance managers setting standards and maturity criteria. Business owners and named data product owners are accountable for defining requirements, prioritising delivery and driving adoption, making it explicit that data products are a business responsibility, not an IT deliverable. Architectural leadership and platform teams ensure coherence and quality, while cross-functional data product teams build pipelines and prepare data up to the product boundary, enabling rapid delivery without fragmenting enterprise consistency.
This separation of responsibilities reduces bottlenecks, makes adoption someone’s explicit job, and allows data products to move faster without undermining trust or governance.
Fund data products beyond the projects that create them
Leaders highlighted a recurring failure mode: data products built for projects lacked funding once delivery ended. One organisation addressed this through graded product levels and central funding for the “last mile” to support products that delivered enterprise value. Without sustainable funding, data products remain temporary assets rather than enduring capabilities.
Protect momentum by limiting collaboration
Over-collaboration surfaced as a drag on speed. Leaders warned against striving for early perfection or one-size-fits-all solutions, advocating instead for incremental delivery and visible progress. Momentum sustains sponsorship. Excessive alignment destroys it.
Become a DataIQ client to access the full set of learnings and many more exclusive insight pieces.
