What makes a high-performance data team

Andreas Bartsch, Head of Innovation and Services at PBT Group

Every organisation wants better decisions to gain a competitive advantage. Yet the path to improved decision-making is often blocked by familiar obstacles. These include fragmented data, unexpectedly escalating cloud costs, and dashboards that keep growing without adding value. Teams chase the “next big use case” while foundational issues remain unresolved.

So, what actually makes a high-performance data team? The honest answer is that it depends. But it depends on specific things.

Start with data maturity

Not every organisation is playing the same game. Some are still building their first consolidated data platform. Their focus is on architecture, modelling, and engineering. Others are modernising legacy estates, migrating to the cloud, and optimising cost and performance. More mature organisations are focused on advanced analytics, AI enablement, and extracting measurable return on their data investments.

A high-performance team looks different at each stage.

Early in the journey, you need a strong data strategy, architecture, modelling, and engineering. You are laying foundations. As the ecosystem matures, cloud expertise, automation, and optimisation become critical. In more advanced environments, governance, data science, machine learning engineering, and AI capability become essential.

The mistake many organisations make is building the wrong team for their level of maturity. High performance is contextual. It is aligned to where you are and where you are going.

Generalists outperform narrow specialists

For many years, data roles have become increasingly specialised. For example, you can have the likes of data modellers, data engineers, data analysts, data scientists, and machine learning engineers, each operating in their own lane.

In practice, high-performance teams are moving back toward multi-skilled data specialists.

My view is that small, multidisciplinary teams outperform large, siloed ones. Yes, specific roles require a strong and relevant academic foundation, particularly in data science. But technical depth alone is not enough. High-performance teams think in terms of outcomes, not outputs. They focus on data products that solve business problems. They understand the industry context in which they operate.

Equally important, they can communicate. Fluency in SQL is valuable. Fluency in English is essential.

Foundation before AI

AI remains a strong focus area. Many organisations are eager to enable AI, but very few pause to assess whether their data ecosystem is ready.

A high-performance data team understands that AI success depends on data quality, governance, lineage, and architecture. Without these, AI simply amplifies inconsistency.

This is where a data ecosystem assessment becomes critical. Before adding new tools, teams should evaluate opportunities for optimisation, automation gaps, cost drivers, and process inefficiencies. Cloud data migration must not only modernise infrastructure but also improve data quality and operational control.

AI enablement is not about finding a clever use case. It is about ensuring that the data platform supports repeatable, governed, value-generating models.

Teams that understand this sequence avoid costly detours.

Structure matters

Over time, organisations tend to oscillate between centralised and distributed data structures.

A fully centralised model promotes consistency and best practice. However, it can become a bottleneck as demand increases. A fully distributed model increases domain expertise and responsiveness but risks duplication and inconsistent definitions.

Most high-performing organisations now adopt a hybrid approach.

Standards, governance, and architectural principles remain centrally defined. Data specialists are embedded within functional teams, working alongside business experts, often guided by a product owner who ensures priorities align with business value.

This model balances agility with control. It allows domain teams to move quickly while maintaining enterprise coherence. It also strengthens accountability for data products rather than isolating responsibility within IT.

Leadership and cadence

Technology alone does not create high performance. Leadership does. Executive support for data adoption ensures that insights translate into action. A team cadence that delivers predictably builds trust. Mature engineering practices prioritise sustainability over one-off successes.

Latest technologies should be leveraged where appropriate. But best-practice data engineering remains the backbone. Without disciplined design, documentation, testing, and governance, scale becomes fragile.

High-performance teams build systems that last.

The crux of the matter

Ultimately, a high-performance data team is not defined by size or tools.

It is a small group of multi-skilled data specialists who:

• Understand the organisation’s maturity level.
• Focus on outcomes rather than outputs.
• Treat data as a product.
• Build sustainable foundations before chasing hype.
• Question assumptions.
• Translate insight into action.

In a year when many organisations are reassessing cloud costs, reviewing AI investments, and seeking measurable returns, these qualities will separate teams that experiment from those that execute.

High performance is not about doing more with data. It is about doing the right things with discipline, clarity, and measurable business impact.