Top 5 metrics for data centre performance

Posted on in

Data centres are at the heart of the UK’s digital economy. Demand for secure, dependable infrastructure continues to rise across cloud services, financial systems, healthcare, and AI workloads. At the same time, expectations around efficiency, transparency, and sustainability are increasing just as rapidly.

With more than 1.6 GW of rated IT capacity across the data centre sector in Great Britain and electricity at its cleanest ever, forecast to be roughly 125 g CO₂/kWh in 2024, the trade-off between dynamism and sustainability has never been as pressing. Operators need to deliver reliable uptime, reduce their carbon footprint, and comply with a growing number of formal ESG standards.

That’s where benchmarks play a role. These are common, standardised measurements that every operator, customer, and regulator can use to understand how well a facility is operating, the risks involved, and how to improve over time. While individual methodologies vary by operator, the underlying concepts remain universally valid benchmarks for measuring efficiency, resilience, and sustainability.

1. Power usage effectiveness (PUE)

The most widely used standard for data centre energy efficiency is Power Usage Effectiveness (PUE). It presents a clear distinction between energy used to power IT equipment and energy wasted powering supporting systems.

Definition

Power Usage Effectiveness is calculated by dividing total facility energy by IT equipment energy. This represents the overhead required for cooling, lighting, and power conversion. A PUE of 1.0 indicates that all kilowatt-hours entering the facility are used directly by IT equipment, though in practice, some infrastructure overhead is always present.

Benchmarks and context

Worldwide, the average PUE is now approximately 1.56 in 2024. In Europe, the Climate Neutral Data Centre Pact sets clear commitments for 1.3 for cool climates and 1.4 for warm climates, effective from 2025 for new sites and from 2030 for existing locations.

Why it matters

PUE remains a valuable headline metric because it enables a clear discussion on how efficiently a facility converts energy into compute. While it does not account for total IT usage or carbon footprint on its own, it provides indispensable context for energy performance and operational cost management.

2. Water usage effectiveness (WUE)

When it comes to building efficiency, energy gets the spotlight, but water use is gaining visibility as sustainability reporting becomes more prevalent.

Definition

Water Usage Effectiveness (WUE) gauges the litres of water used per kilowatt-hour (kWh) of IT energy usage. This captures the water required to cool and humidify the site.

Benchmarks and context

Facilities using evaporative or water-based cooling typically record higher WUE values than air-cooled systems. The Climate Neutral Data Centre Pact targets 0.4 litres per kWh for cool climates using potable water systems by 2025. In the UK, where water stress varies by region, reducing WUE aligns with local resilience and national sustainability objectives.

Why it matters

Unlike PUE, WUE measures the pressure on natural resources beyond electricity. Integrating water-use efficiency into a broader long-term sustainability strategy ensures that data centre operations remain resilient even as local environmental regulations and resource availability tighten.

3. Carbon usage effectiveness (CUE)

Carbon Usage Effectiveness bridges the energy-efficiency and environmental-impact gap, but the environmental side is weighted more toward emissions than to use.

Definition

Carbon Usage Effectiveness (CUE) measures the kilograms of carbon dioxide emitted per kilowatt-hour of IT energy used. This determines not just how efficiently energy is used, but also how cleanly it is sourced.

Benchmarks and UK context

Across the UK grid in 2024, intensity was approximately 0.124 kg CO₂ per kWh. The Climate Neutral Data Centre Pact targets 100% carbon-free energy by 2030 and therefore reinforces the focus on CUE for Scope 2.

Why it matters

CUE translates engineering to climate impact. It provides transparent and consistent carbon reporting, which is essential for sustainability communications with customers and regulators.

4. Availability and uptime

Efficiency statistics only reveal half the battle. Ultimately, operational integrity and service delivery determine whether a data centre can truly provide service under real-world conditions.

Definition

Data centre uptime refers to the percentage of time a facility is fully operational and powered, while availability measures the proportion of time a system is functionally accessible to end users. While they are often used interchangeably, true availability factors in both uptime and any performance degradation or planned maintenance windows that might affect service delivery.

Industry benchmarks

Tier III facilities target 99.982% availability, allowing for 1.6 hours of annual downtime, while Tier IV standards demand 99.995%, which permits just 26 minutes. Despite high design standards, around 55 per cent of operators have experienced at least one outage in the past three years.

Why it matters

Availability underpins trust. Even brief outages can be highly disruptive to customers running mission-critical workloads. This reliability is the result of rigorous, audited security and operational standards that ensure infrastructure is concurrently maintainable and resilient to disruption.

5. Capacity utilisation and resource planning

Capacity utilisation is a critical operational variable that is often overlooked, but is vital for efficiency, scalability, and ROI.

Definition

It is measured as the percentage of actual output relative to the potential maximum. While often grouped, there is a distinct difference between resource planning and capacity planning. Resource planning manages human and technical assets, while capacity planning focuses on the physical and thermal limits of the facility.

UK capacity context

The colocation capacity in Great Britain was estimated to have grown to around 1.6 GW by 2024. London is at 1,048 MW, while the South East contributes 128 MW and the North West 52 MW.

Why it matters

Capacity utilisation is the bridge between operational efficiency and business strategy. It ensures that infrastructure investments remain resilient while catering to present needs with adequate headroom for growth.

The strategic value of performance data

One measure does not make a data centre. Each of PUE, WUE, CUE, uptime, and capacity utilisation provides a different view of efficiency, resilience, and sustainability. Together, they offer the clarity required to understand how a facility is currently performing and how to improve it tomorrow.

Understanding these metrics is the first step toward building resilient, sustainable infrastructure. At Datum, we move beyond the numbers by providing enterprise-grade environments and audited operational standards designed to support your mission-critical workloads.

Book a tour of our secure facilities in Farnborough or Manchester to see our infrastructure in person, or speak with our expert team to discuss your specific colocation requirements.

Back to Insights