There is a question that comes up repeatedly among Australian solar owners, and it almost always sounds the same. Something like: “My system seems to be working, but I am not sure if it is working as well as it should be”. Or: “My bills went down after installation, but I have no idea if that is the right amount”. Or, most honestly:”I check the app occasionally, see some numbers, and have no real way of knowing if those numbers are good or not”.
That uncertainty is not irrational. It is the correct response to a genuine gap in how most solar systems are monitored.
An Ausgrid study across 8,000 solar systems found that 51.8% were not reaching optimal generation. More than half were generating power, appearing to function normally, and quietly losing hundreds of dollars a year in output their owners never knew was missing.
The monitoring app was showing numbers. The bill was lower than before solar. Everything looked fine.
This is the invisible underperformance problem in Australian solar, and it is far more common than the industry tends to acknowledge. This post is about how to know whether your system is one of the majority that has a problem, or one of the minority that genuinely does not.
The gap your monitoring app cannot close
Your monitoring app tells you what your system is generating. It does not tell you what your system should be generating. Those are two completely different pieces of information, and the second one is the only one that tells you whether something is wrong.
This distinction matters more than it sounds. A system generating 18 kilowatt-hours on a clear October day might be performing excellently for a 5-kilowatt system in Hobart, adequately for a 6.6-kilowatt system in Melbourne, or badly for an 8-kilowatt system in Brisbane. The number itself is meaningless without a reference point. And the reference point most monitoring apps provide is yesterday, or last week, or last month. They compare your output to your own history. They do not compare it to what your system should theoretically produce given your location, your system size, your panel orientation, and today’s actual weather.
That is the gap, and it is the gap that allows a poorly configured inverter, a failing string, a degraded set of panels, or a rooftop isolator slowly deteriorating in the weather to go undetected for months or years.
Two numbers worth knowing
These are the two other metrics that are more useful for diagnosing whether your system is performing the way it should be:
- Performance ratio: This measures how efficiently your system converts available solar energy into electricity, accounting for real-world losses from heat, wiring, inverter inefficiency, and other factors. A well-designed and properly functioning system should achieve a performance ratio between 0.75 and 0.85. On a clear day, if you have access to a monitoring platform that calculates this figure, anything consistently below 0.75 is worth investigating.
- Specific yield: This measures energy production per kilowatt of installed capacity, expressed as kilowatt-hours per kilowatt-peak per year. For a residential system with reasonable orientation and minimal shading, a specific yield in the range of 950-1,300 kilowatt-hours per kilowatt-peak per year is a realistic expectation, depending on your state. For a 6.6. kilowatt system in Sydney, which means annual output in the range of 6,270 to 8,580 kilowatt-hours. If your system has been producing significantly below the lower end of that range since installation and nobody has ever investigated why, that is a question worth asking.
Neither metric requires specialist equipment to use as a rough diagnostic. Both require knowing your system size, your location, and your annual generation figure from your monitoring app or electricity bill.
The DIY method that actually works
One approach discussed in solar communities is the same-date comparison: go back through your monitoring app to the same date last year and the year before, find a clear day in each, note the daily total generation, and compare against a recent clear day in the same seasonal window.
This is actually sound. It is the manual version of what professional monitoring platforms do automatically, and, done carefully, it can surface genuine problems. A few things to make it work better:
- Find clear days specifically, not just any day: Cloud cover is the biggest variable in daily generation, and comparing a partly cloudy day this year against a clear day last year tells you nothing useful. Most monitoring apps allow you to see historical daily totals; look for days where the generation curve is smooth and peaked, which indicates clear sky conditions.
- Use multiple comparison points: One anomalous day in either direction can mislead you. 3 or 4 clear days from the same seasonal window across 2 or 3 years gives you a more reliable picture.
- Apply a degradation adjustment: Solar panels lose output capacity gradually over time, typically between 0.3 and 0.8% per year. A working assumption of roughly 1% per year of age is reasonable for most residential systems. A 5-year-old system should be producing around 95% of its original rated output on an equivalent clear day. A 10-year-old system, around 90%. If your system is 3 years old, you would expect roughly 3% lower output compared to its first year under identical conditions. A gap of 15% points to something beyond normal ageing.
The honest limitation of this method is that it catches step-change problems well but struggles with slow drift. If your system has been losing 2-3% of output per year rather than the expected 1%, the decline can feel gradual enough to seem normal year on year. You would only notice it clearly if you compared the current output against a baseline from 4-5 years ago. Most homeowners do not think to do that until the loss has become significant.
This is what automated monitoring solves.
What automated monitoring does that manual checking cannot
Solar Analytics, the most established Australian residential monitoring platform, uses satellite data and Bureau of Meteorology irradiance readings to calculate what your system should be producing at your specific location on any given day. It then compares that figure against what your system is actually producing. The difference is not between you and your own history. It is between you and a physics-based model of what your panels should theoretically generate, given today’s actual conditions at your address.
It also compares your generation against nearby systems in the same area. If all systems in a suburb are generating 20% below the model on the same day due to atmospheric haze or smoke, the platform accounts for that and does not raise a false alarm. If your system is generating 20% below the model while your neighbour’s systems are performing normally, that is flagged as a potential fault.
Research from the University of Technology Sydney and UNSW has validated this approach on a large scale. Working across 1,089 Australian PV systems, the team found that automated fault detection using only standard inverter data, without additional hardware or sensors, achieved 92% accuracy in classifying underperformance events. Their threshold for flagging major underperformance: a measured-to-expected ratio below 60% for 3 consecutive days.
That threshold is the clinically clear case. The more common and more financially costly situation is the system sitting at 75085% of expected output, month after month, for reasons that would be straightforward to fix if only someone had noticed.
The faults hiding in plain sight
Understanding what causes silent underperformance makes the diagnostic conversation more concrete. These are the most common issues in residential systems that produce no obvious fault signal but do produce measurable output losses:
- Incorrectly configured inverter settings: These are more common than the industry acknowledges, particularly in systems installed during the rapid-growth period of 2010-2016, when installer quality varied considerably. Without a performance baseline to compare against, an owner has no way of knowing the system has been underdelivering since the day it was switched on.
- Roof isolator deterioration: This is one of the leading causes of system failure in Australia. Australian regulations require a rooftop DC isolator, a plastic enclosure with an isolation switch, to be installed on every residential system. Sitting on a rooftop exposed to heat, UV, rain, and temperature cycling for a decade or more, these components are prone to water ingress, overheating, and internal failure. The result can range from gradual output reduction to complete system failure, with no dashboard indicator that anything is wrong.
- String-level underperformance: This is usually due to shading, soiling, or bird activity. A single panel producing below its rated output in a standard string-inverter system reduces the output of every panel in that string, not just the affected one. A bird nest under the array, a not spot caused by a cracked cell, or a section of panels shaded by a tree that has grown since installation can each knock a full string’s output down significantly, with no visible indication on the inverter.
- Generation clipping: This is usually from an inverter that is undersized relative to the panel array. If your system was expanded with additional panels after initial installation without a corresponding inverter upgrade, the inverter may be clipping peak production every clear day. If the direct-current to alternating-current ratio of your system exceeds around 1.3, clipping losses are likely affecting your output. This is a design problem rather than a fault in the traditional sense, but it produces the same result: output consistently below what the panels are capable of.
- Gradual inverter degradation: Inverters do not always stop working clearly. They can degrade over time, producing output that is measurably lower than a new inverter of the same specifications under the same conditions, without any error code or visible indication that performance has declined. Continuous monitoring is the most reliable way to detect this kind of gradual efficiency loss.
What to do with this information
If you have been operating on the assumption that your system is fine because the app shows generation and the inverter light is green, the Ausgrid data suggests that assumption has roughly even odds of being wrong.
The same-date comparison method described earlier is a reasonable starting point if you want to do a quick self-check before calling anyone. Pull up your monitoring app, find 3 clear days from the same season in previous years, compare against a recent equivalent clear day, apply the 1% per year degradation expectation, and see how the numbers look. If the difference is within a few per cent of what ageing would predict, your system is probably performing close to normally. If it is larger, you have a starting point for a conversation with a qualified installer.
If your system is more than 3 years old and has never had a professional inspection, that check is worth doing regardless of what the app shows. The faults most likely to be losing you money are the ones that produce no visible signal: the misconfigured setting, the deteriorating isolator, and the underperforming string that blends into the overall generation figure.
Knowing what your system should be producing, and comparing that against what it is actually producing, is not an advanced technical exercise. It is the basic question that more than half of Australian solar owners have never had a clear answer to.
Not sure where your system sits against the numbers in this post? Solar Service Guys carry out professional solar health checks covering inverter diagnostics, string testing, and panel inspection. If your system has been quietly underperforming, a service check will tell you exactly where the losses are coming from.
