Why Your Solar Installer's Estimate and Your Actual Production Don't Match
Last updated: 2026-04-08 · Solar Benchmark
Why Your Solar Installer's Estimate and Your Actual Production Don't Match
Installer sales estimates average 10-15% above actual first-year production, according to NREL analysis. The gap comes from optimistic shading assumptions, TMY weather data that does not reflect your actual year, and modeling choices that favor larger production numbers at the point of sale.
Head-to-Head Comparison Table
| Factor | Installer Sales Estimate | Physics-Based Benchmark |
|---|---|---|
| Weather data source | TMY (30-year average) or proprietary dataset | ERA5 hourly reanalysis, actual calendar year |
| Shading modeling | Simplified or optimistic | Hourly sun angle geometry with horizon data |
| Soiling assumption | Often omitted or minimal | Regional soiling rates applied seasonally |
| Purpose | Support the sale of a solar system | Diagnose whether the installed system performs correctly |
| Accuracy vs measured | 10-38% variance depending on tool | 5-7% monthly deviation |
| When to use | Before installation, for system sizing | After installation, for ongoing performance verification |
What Each Approach Measures
Installer sales estimates are produced using PVWatts, Aurora Solar, Helioscope, or similar tools. PVWatts applies a fixed performance ratio of 0.78 and uses TMY weather data, producing approximately 38% variance from actual measured performance in NREL validation studies. Aurora and Helioscope use more detailed shading models and proprietary datasets that narrow variance to roughly 10-20%, but both still use weather data that represents a typical year rather than the actual year your system will operate in.
Sales estimates are also shaped by the sales context. The estimate is a document that supports a purchasing decision. Installers using the same modeling tools can produce meaningfully different estimates by adjusting shading assumptions, soiling loss inputs, and performance ratio settings. These inputs are rarely disclosed in the customer-facing estimate document.
Physics-based benchmarks use ERA5 atmospheric reanalysis, which records actual hourly irradiance at your coordinates for the specific calendar period being evaluated. The model applies your panel's CEC-certified efficiency and temperature coefficient, your confirmed tilt and azimuth, regional soiling rates, and inverter efficiency curves. The output is what your system should have produced given what the atmosphere actually delivered to your roof. See /resources/methodology for full methodology documentation.
Key Differences
Weather data. Installer estimates use TMY data representing a synthetic average year. Any real year can differ from TMY by 5-15% in irradiance. If you had an unusually cloudy first year, your production will be below the TMY-based estimate regardless of whether your system is working correctly. The benchmark uses actual weather, so it separates weather effects from system performance.
Shading. Aurora and Helioscope use satellite and LiDAR data to model shading from nearby structures. In practice, shading is frequently underestimated because the satellite imagery used at the time of sale may not reflect vegetation growth, new construction, or seasonal shading patterns. PVWatts cannot model time-of-day shading at all.
Soiling. Many installer estimates apply zero or near-zero soiling loss. In reality, dust, pollen, bird droppings, and particulate matter reduce output by 2-6% annually in most US climates before cleaning, with higher losses in arid Southwest regions. NREL data supports regional soiling loss rates that physics-based models apply by default.
Purpose and incentives. The installer's estimate is a sales document. Its purpose is to help a homeowner feel confident in the payback calculation for a significant purchase. A physics-based benchmark produced after installation has no stake in whether the number comes out high or low. It reflects what the hardware and weather allow.
Legal standing. Most installer contracts include explicit disclaimer language stating that production estimates are not guarantees. Solar leases and power purchase agreements (PPAs) sometimes include minimum production guarantees that trigger compensation if output falls below a threshold. Cash purchase and loan contracts almost never include production guarantees. The estimate in your sales proposal is not a contractual commitment in most cases.
When Each Option Makes Sense
The installer's sales estimate is useful for:
- Comparing system size options before purchase (8 kW vs 10 kW)
- Estimating payback period and ROI as part of a purchasing decision
- Understanding which roof planes the installer plans to use and why
A physics-based benchmark is the right tool for:
- Verifying whether your installed system performs at the level the hardware allows
- Distinguishing between a bad-weather year and a failing system
- Building documentation for a warranty claim or installer dispute
- Annual performance tracking against a consistent, repeatable standard
Frequently Asked Questions
Is my solar installer liable if production is lower than estimated?
In most cash purchase and loan contracts, no. Standard installer contracts include disclaimer language stating that production estimates are projections based on modeling assumptions and are not performance guarantees. If you have a solar lease or PPA with an explicit production guarantee clause, the agreement's terms determine your recourse if output falls below the guaranteed threshold. Review your contract for guarantee language before assuming any protection applies. Source: SEIA residential solar contract guidance.
How can I tell whether my system is underperforming versus just having a bad-weather year?
Compare your actual production against a physics-based benchmark calculated using ERA5 weather data for your specific location and the calendar period you are evaluating. ERA5 reflects the actual irradiance your roof received, not a TMY average. If your system produced 88% of the physics-based expected value, the 12% gap is a system performance issue, not a weather issue, because the weather is already baked into the expected value. NREL validation places this model's accuracy at 5-7% monthly deviation, so a gap above 10% reliably indicates a real performance problem.
Why do Aurora and Helioscope estimates still miss actual production?
Both tools use more accurate shading models than PVWatts, but they still rely on pre-installation satellite and LiDAR data that may not reflect actual conditions, and they use weather datasets that represent typical years rather than actual calendar-year irradiance. They also depend on accurate input of tilt, azimuth, and module orientation, which can contain installer entry errors. Even with these tools, variance of 10-20% against actual measured performance is common.
What production shortfall is worth investigating?
Use a physics-based benchmark as the reference, not the installer's estimate. A system performing 5-7% below the benchmark falls within normal model uncertainty. A gap of 10-15% warrants investigation, especially if it persists across multiple months. A gap above 15% against a physics benchmark indicates a measurable system problem: shading that was not modeled, failed equipment, soiling, or a commissioning issue. These thresholds align with NREL performance monitoring guidance and IEC 61724 standards.
Data: pvlib physics modeling + Open-Meteo ERA5 weather data | Last updated: 2026-04-08 | Solar Benchmark