Decarbonarma Heat pumps HeatpumpMonitor.org
Buy Me a Coffee QR code ☕ Scan to support

Real-world data

What 600+ heat pumps actually tell us

HeatpumpMonitor.org is the UK's most rigorous open dataset on heat pump performance — built by owners, for owners. Here's what the numbers reveal about efficiency, flow temperatures, and what good design really looks like.

Data review Source: heatpumpmonitor.org & OpenEnergyMonitor Updated Nov 2025
639 live systems uploading data
448 full MID-metered systems
3.86 average SPF H4 (air source, no cooling)
213 systems with a full year of data

What is HeatpumpMonitor.org?

HeatpumpMonitor.org is an open-source community initiative run under the OpenEnergyMonitor umbrella. Heat pump owners — many of them among the UK's most technically rigorous installers and enthusiasts — connect monitoring hardware to their systems and upload live performance data. The result is a growing, publicly searchable dataset covering everything from seasonal performance factors to flow temperatures on the coldest recorded day.

The dataset is genuinely useful for anyone buying, installing, or simply curious about heat pumps. It answers questions that manufacturer specifications simply can't: how does a particular model behave in mild weather? What does it actually output during a defrost cycle? Is this system's performance typical or outlying?

The central finding: low flow temperature is everything

The single strongest predictor of annual performance (R² = 0.55) across all 600+ systems is the difference between the average flow temperature and the outside temperature. The lower this gap, the higher the SPF. This confirms basic thermodynamic theory — but it's now proven against hundreds of real UK homes.

In practice, the data shows a clear gradient. Systems running with a weighted average flow temperature of around 30°C achieve an SPF of roughly 4.4. At 35°C average flow, that drops to around 3.9. At 40°C, it falls to about 3.4. Every extra degree of flow temperature costs efficiency.

Weighted avg flow temp Typical SPF H4 Assessment
30°C~4.4Excellent — optimally designed system
35°C~3.9Good — well-tuned radiator system
40°C~3.4Moderate — room for improvement
45°C+<3.0Poor — under-sized emitters or wrong settings

The flow temperature myth: systems run cooler than designed

One of the most striking findings is that most systems actually run at lower flow temperatures than their heat loss calculations predicted. Systems designed for a 45°C flow temperature typically run at around 38°C on their coldest day. Those designed for 50°C typically achieve around 40°C in practice.

"Systems with an SPF of 4.0 have average flow temperatures on the coldest days of 36.6°C ± 5.5°C — well below the 45–50°C rules of thumb still commonly used in design."

The implication is significant: heat loss calculations, particularly for pre-2000 properties, frequently overestimate how much heat a home actually needs — especially when fabric improvements like loft insulation and double glazing are properly accounted for. When weather compensation is tuned to the real demand rather than the calculation, flow temperatures fall and efficiency rises.

Design recommendations from the data

The OpenEnergyMonitor team distils the dataset into clear design guidance. The key thresholds are:

What this means for your installation

The dataset validates what the best installers already know: brand choice matters far less than system design. A well-designed, properly commissioned heat pump running at 35°C flow in a home with adequately sized radiators will comprehensively outperform a premium unit running at 50°C in a poorly designed system. The data shows SPF H4 values ranging from below 2 to above 5 — almost entirely explained by system design and commissioning, not hardware.

For those buying now: insist on a proper heat loss calculation, ask what design flow temperature your installer is targeting, and make sure weather compensation will be set up and tuned at commissioning. These decisions matter far more than the brand badge on the outdoor unit.

Data and analysis sourced from HeatpumpMonitor.org and the OpenEnergyMonitor documentation. Statistics reflect the period November 2024 – November 2025. The dataset is open-source and publicly searchable — you can explore individual systems, compare models, and access raw performance data directly on the site.
Free tools for this topic
🏠 Heat Loss Estimator →
Related heat pump reading