How do I measure what my heat pump can actually deliver, rather than trusting the headline capacity?

The headline number is real but it is quoted at a specific operating point, and neither of the two variables that define that point is the one you care about.

Capacity is a function of outdoor temperature and flow temperature, and it moves a lot. Colder outside means less heat available to pick up, so capacity falls. Hotter flow temperature means a bigger lift for the compressor, so capacity falls again and efficiency falls with it. A unit rated at a mild outdoor temperature and a modest flow temperature can deliver appreciably less at your actual design condition — a cold day with the flow temperature high enough to make your radiators work.

So the number you want is capacity at your design outdoor temperature and your actual flow temperature, and it is in the manufacturer's full technical data rather than the brochure. Look for the capacity tables that vary both axes. The brochure figure and the table figure for your conditions are frequently not close.

Measuring the real thing. Heat output is flow rate multiplied by the temperature difference across the unit, multiplied by the specific heat capacity of the fluid. In practice:

• You need a flow meter. This is the one measurement you cannot infer, and without it everything else is arithmetic on a guess. A heat meter that does the whole calculation is the tidiest option; a flow meter plus your existing temperature sensors also works.
• Watch out for glycol. If the system has antifreeze in it, the specific heat capacity is lower than water's and the fluid is more viscous, so you get less heat for the same flow and delta. Ignoring it overstates output by a meaningful margin.
• Measure at steady state. During a defrost cycle output is negative — the unit is taking heat out of your system to melt the ice. Averaging over a period that includes defrosts is the correct way to judge real-world seasonal performance, but for "can it hit its rated output" you want a clean steady run.

Then the comparison that answers your actual question. Take a genuinely cold period and record: outdoor temperature, flow temperature, measured heat output, electrical input.

• If measured output is close to the table figure for those conditions, the unit is fine and your problem is elsewhere — most likely emitters, which is the next paragraph.
• If it is well below, and the delta across the unit is small, you probably have a flow rate problem rather than a compressor problem: undersized pipework, a partially closed valve, a pump on too low a setting, or a clogged strainer. This is common and fixable.
• If output is low and the unit is cycling rather than running continuously, it is a controls or sizing issue.

On the emitters, because it is the likeliest answer to "the house did not hold temperature": a radiator's output depends on the difference between its surface and the room, so a radiator sized for a high-temperature boiler delivers substantially less at heat pump flow temperatures. The unit can be delivering exactly its rated output into a system that cannot get that heat into the rooms. You would see this as the unit running continuously, hitting its flow target, and the rooms still being cool — and no amount of measuring the heat pump will reveal it. Compare each room's radiator output at your actual flow temperature against that room's heat loss.

One thing to record alongside everything else: whether the immersion or backup heater came on.

A lot of systems will quietly bring in electric backup on the coldest days to make up a shortfall. If that happened, your "the house did not quite hold temperature" may have been considerably more expensive than you realise, and the heat output you measure includes heat the heat pump did not produce.

It is usually visible as a step change in electrical input with no corresponding change in the compressor's behaviour.

Strong agreement on the emitters, and there is a cheap diagnostic that separates the two cases in one evening.

On a cold day, deliberately raise the flow temperature well above your normal curve — as high as the unit will go. If the house then warms up properly, the heat pump can deliver and your emitters are undersized for the flow temperature you normally run. If it still cannot hold temperature at maximum flow, the unit or the flow rate is the limit.

It is inefficient and you would not run like that, but it is an hour and it tells you which half of the system to spend money on. Note the electrical consumption while you do it, because that is also a good illustration of why you do not run like that.

One caution: do not conclude anything from a single mild day. The test only means something near your design condition.

Practical note on the flow meter, since it is the purchase that unlocks all of this: fit it somewhere you can actually read it and where the pipe run is straight for a reasonable distance either side.

Most of them specify a minimum straight length upstream and downstream, and installers routinely put them in an elbow because it was convenient. The reading is then wrong in a way you cannot detect from the number itself, and you spend a season drawing conclusions from bad data.

Same for the temperature sensors — good thermal contact and insulated over the top, or the flow reading drifts towards room temperature and your delta is understated.