“We’re building a lakehouse on Databricks” isn’t a vision on its own, it’s a technology choice. Successful platforms articulate their purpose in terms of business value: enabling faster decision-making, reducing regulatory risk, accelerating product innovation or lowering operational costs.

This vision is a catalyst for intelligent adaptation and evolution. When new capabilities are proposed, the question isn’t “can we build this?” but “does this move us closer to the outcomes we are working to enable?” When measuring success, you’re tracking business impact, not just technical milestones.

Without this level of clarity and consistency, platforms are often technology exercises disconnected from the problems they’re meant to solve. With it they become strategic assets with a clear line of sight to organisational goals.

Platform teams often turn to activity metrics as their first port of call. Some usual suspects: data sources onboarded, pipelines deployed, tables in the catalog. Useful numbers but not tightly coupled to value.

Product thinking demands outcome metrics: time-to-insight for analytics teams, percentage of self-service requests vs. manual fulfilment, reduction in shadow IT systems or direct business impact like revenue change influenced by platform-enabled features.

Value-based metrics fundamentally change prioritisation calls. If you’re measuring tickets closed, you optimise for throughput. If you’re measuring time saved for data consumers, you’re now prioritising self-service tooling, documentation and automation, the capabilities that scale and make your users happier to work with your platform.

Data platforms serve multiple dynamic user groups: data engineers building pipelines, analysts exploring datasets, ML engineers training models, business users consuming reports and governance teams ensuring compliance. Each has different workflows, skill levels, values and needs.

UCD means understanding these personas deeply and designing for their job-to-be-done, not stopping at architectural elegance. It means intuitive interfaces for common tasks, clear enabling documentation written for practitioners and reducing friction at every step.

Platforms designed primarily for infrastructure teams often become power tools that rely on expert knowledge. Platforms designed for users become adoption and collaboration engines that scale beyond the initial technical audience.

Centralised gatekeeping doesn’t scale gracefully. When every data request requires a ticket to the platform team the team serves as a bottleneck, users get frustrated and shadow IT proliferates as a workaround.

Self-service means enabling users to provision data products, request access, explore datasets and build pipelines without manual intervention or dependence. However this is paired with maintaining appropriate guardrails through automation, templates and embedded governance rules.

Counterintuitively, self-service doesn’t diminish the platform team’s importance, it elevates it. Instead of executing individual requests the team becomes focused on building capabilities that enable hundreds/ thousands of requests, shifting from operators to enablers.

Governance is where data platforms diverge from generic product platforms. Governance isn’t optional but users won’t enthusiastically adopt it unless it’s frictionless and valuable.

The best platforms make governance something users want to engage with. Automated lineage shows users where data comes from and where it’s used. Built-in quality metrics help users know what they can trust. Access transparency reduces security anxiety. Policy-as-code applies governance rules automatically, not through manual review gates.

When governance is bolted on as an afterthought, it becomes a compliance tax that users try to avoid. But when data platform teams design it as a platform capability from day one it becomes a competitive advantage.

Waiting until it’s perfect kills data platforms before they launch. Typically by the time you’ve built everything the requirements have changed, the technology has evolved and users have lost patience.

Product-thinking platforms ship incrementally. They start with one high-value use case, prove the model then expand. Each phase delivers measurable value and generates feedback that shapes the next step forward. This approach de-risks investment, accelerates time-to-value and builds momentum and morale through visible successes.

Incremental delivery also embeds natural feedback loops that protect against building the wrong thing. You learn what users truly need continually and directly not what they said they needed 18 months ago in a requirements doc.

Data platforms as traditional IT projects operate through handoffs: requirements → design → build → deploy → live service. Each function working in isolation, passing artifacts over the wall and co-existing but rarely truly collaborating.

Data Platform as a Product (DPaaP) operates with collaboration at its core. Multidisciplinary teams own outcomes end-to-end. Data engineers, platform engineers, governance specialists and product owners work together in sprints, demo regularly and iterate based on direct feedback.

This is a nod and recognition that platform challenges are rarely purely technical. Adoption is a people problem. Governance is a workflow problem. Performance is a user experience problem. Cross-functional teams can address these far more holistically, delivering value end to end without dependency in a way that siloed functional teams just can’t.

How do you know if your platform is succeeding? Not by asking the steering committee or the platform team. By listening to your users, observing their behaviour and analysing their interactions with your data platform.

Effective feedback loops include usage telemetry (what features are used, where users get stuck), regular user research (shadowing sessions, surveys, office hours), a transparent roadmap that invites external input and celebrating user success stories to foster cultural adoption.

Without these mechanisms, platform teams operate in the dark, building what they think users need, rather than responding to real pain points or opportunities. With them, platforms evolve based on evidence and proven demand.

No data platform exists in isolation, and yours is no exception. The best integrate with existing tools and position themselves as the connective tissue across the entire data ecosystem.

This means interoperability with BI platforms, ML frameworks, orchestration tools and legacy systems. It means cloud-native architecture that leverages managed services rather than reinventing infrastructure. It means open interfaces (APIs, SDKs, webhooks) that let advanced users extend the platform and run their own experiments to accelerate learning across the board.

Organisations rarely have the luxury of greenfield environments. Platforms that embrace the existing ecosystem accelerate adoption because users don’t have to abandon familiar tools, jump from burning platforms or force disruptive migrations.

Treating your data platform as a product fundamentally shifts how you measure success, engage users and evolve capabilities. It changes accountability, prioritisation and daily operations from reactive project delivery to continuous value creation.

Don’t feel pressured to implement all ten features overnight. Product thinking is a journey that begins with small, deliberate shifts and at its core is around maximising the surface area that learning and collaboration can thrive on.

Start by naming a Product Owner with clear accountability. Choose one user segment and focus on delivering measurable value for them first. Shift one metric from activity to outcomes and let that drive your next decisions.

Each step builds a little more momentum and each success creates the fuel for the next capability to grow from.

By combining solid technical data platform skills, adaptive delivery practices and user-centred product thinking data teams can more easily close the gap between technically impressive platforms and truly valuable ones that users love. Turning your data platform from a technical masterpiece to a strategic enabler and capability-building product that empowers users, accelerates innovation and builds lasting trust across the organisation.