What Belongs in a Solar O&M Contract: Availability, Performance Ratio and the Clauses That Matter

A solar array is the only major capital asset a business will buy that appears to need nothing.

There is no noise, no fuel, no moving part to watch, and no obvious sign when something is wrong. Panels sit on a roof looking exactly as they did on the day of commissioning, whether they are producing 100% of expectation or 68%. That is precisely why solar operations and maintenance is undersold, underpriced and, in a great many installations, quietly absent.

The consequence is a well-documented pattern: a system performs as promised for the first year, then declines on a curve nobody is watching. Soiling accumulates, a string goes offline, an inverter derates in the heat, a combiner fuse blows. Nothing alarms because nothing is monitored. Two years later the owner notices the generator running more than it used to and assumes the panels have worn out.

A solar O&M contract is what prevents that. This article covers what belongs in one, the difference between the two guarantees that matter, what the service should actually cost, and the clauses that decide whether you have bought a service or a subscription to an annual visit.

Availability and Performance Ratio Are Not the Same Guarantee

Almost every dispute over a solar O&M contract traces back to confusion between these two metrics. They measure different things, they protect against different failures, and a contract with only one of them has a hole in it.

Availability measures whether the system was capable of operating. It is typically expressed as the percentage of daylight hours in which the plant, or a defined portion of it, was online and able to produce. Industry practice puts availability guarantees somewhere in the range of 97% to 99% per year.

Availability is the right measure of the service provider's responsiveness. If an inverter fails on a Monday and is back on Wednesday, that downtime is squarely within the contractor's control, and the guarantee holds them to it. What availability does not capture is a plant that is fully online and producing badly.

Performance ratio measures how much of the energy available in the sunlight that actually fell on the array was converted into delivered electricity. It captures soiling, module degradation, string mismatch, cable losses, inverter efficiency, temperature derating and shading. A well-maintained system typically runs a performance ratio somewhere between 75% and 85%, and a ratio that falls below roughly 70% is a clear signal that something needs physical inspection.

Performance ratio is the metric that catches the failure mode owners actually suffer from, because the most common form of underperformance is not an outage. It is a system that is fully available and dirty.

The contract should carry both. Availability alone lets a contractor meet every obligation on a plant losing 20% to soiling. Performance ratio alone penalises the contractor for weather and for module degradation that is nobody's fault, unless the calculation is properly normalised for irradiance and temperature, which it must be.

Two details make the difference between a guarantee and a decoration. First, the guarantee must specify a remedy: a service credit, a liquidated damages calculation, or compensation for lost energy at a defined rate. A guarantee with no consequence is a target. Second, the measurement basis must be defined precisely, including the irradiance sensor used, how it is calibrated and how often, and which exclusions apply. Disputes are almost always about measurement, not about performance.

What the Scope Should Actually Contain

A serious O&M scope splits into three activities that are frequently conflated and priced as one.

Preventive maintenance is scheduled work: array cleaning at an interval matched to the site's soiling rate, visual inspection of modules and mounting, thermographic inspection of modules and DC terminations, torque verification on electrical connections, inverter servicing including filter cleaning and firmware, combiner and protection device checks, string current measurement, earthing continuity and resistance testing, and vegetation or shading management. Annual thermal imaging is one of the highest-value items in the list, because loose DC terminations are both a production loss and a fire risk.

Corrective maintenance is fault response: diagnosis, repair, replacement and the associated logistics. This is where response times and spares holding matter.

Monitoring and reporting is the activity that makes the other two possible. Without continuous monitoring, preventive maintenance is guesswork and corrective maintenance begins when someone happens to notice.

Cleaning frequency deserves particular attention in this region, because it is the item most often specified generically and most often wrong. Soiling loss in dusty, low-rainfall conditions is severe, and the correct interval is a function of the site rather than of the calendar. Our guide to solar panel maintenance in dusty climates sets out how to establish a realistic interval, and a contract that specifies "cleaning as required" without defining the trigger has specified nothing.

Reviewing an O&M proposal or specifying one for tender? Our renewable energy team writes scopes that define the trigger, the measurement and the remedy rather than the frequency alone.

Availability and performance ratio measure different failures. A plant can be fully available and losing a fifth of its output to soiling. Source: MIMAH engineering analysis; performance ratio and availability bands per industry O&M practice.

What It Should Cost

Owners consistently underestimate this line, and contractors compete on it, which produces a market where the cheapest quotation is frequently the one that cannot deliver the scope it describes.

Published benchmarks put commercial solar O&M at roughly 0.5% to 1.5% of system capital cost per year, or on a capacity basis somewhere in a planning range of a few dollars per kilowatt peak per year, with the figure varying substantially by system size, site complexity, accessibility and the depth of scope included. Mature European markets with clean environments and easy access sit at the low end of the range; complex or remote sites with heavy soiling sit well above it. NREL's assessment of solar PV plant O&M budgeting practices sets out how widely these figures vary with scope definition, which is precisely why comparing two quotations on price alone is meaningless.

The more useful way to evaluate cost is against what underperformance costs. A more expensive contract that holds a plant above 99% availability can be substantially more profitable than a cheap one that lets it sit at 94%, because the lost energy exceeds the price difference several times over. On a system of any size, a few percentage points of annual production is worth more than the entire O&M fee.

This is the arithmetic that owners should insist on seeing. Take the plant's expected annual production, apply the value of a kilowatt-hour at the site, whether that is the avoided diesel cost or the PPA tariff, and calculate what 5% of underperformance costs per year. In most commercial installations in this region, where the alternative to solar is diesel at a substantial multiple of the solar cost, that number dwarfs the maintenance contract.

Note also who bears the cost of underperformance depending on the financing structure. Under a power purchase agreement the developer carries it, which is why PPA providers maintain their systems properly: their revenue depends on it. Under a direct purchase or a lease, the owner carries it entirely, which is exactly when a well-specified O&M contract matters most. Our guide to commercial solar financing covers how that risk allocation differs between structures.

The Clauses That Decide Whether It Works

Beyond scope and price, a small number of provisions separate a contract that delivers from one that produces an annual visit and a certificate.

Response times, tiered by severity. A total plant outage and a single string fault are not the same event and should not carry the same response obligation. Define severity tiers with distinct response and resolution times, and define whether the clock runs on calendar hours or working hours, because that distinction is worth days.

Spares holding. Who holds critical spares, where, and which ones? An inverter with a twelve-week import lead time makes the response time clause meaningless. The contract should specify what is held locally, and for critical installations a spare inverter on site is cheaper than the production lost while one is shipped.

Monitoring platform and data ownership. The owner should have direct access to the monitoring platform, not a monthly summary produced by the contractor. And the data belongs to the owner. Contracts where the monitoring system is proprietary to the contractor create a switching cost that is entirely artificial and entirely deliberate.

Reporting content and cadence. Monthly reports should contain actual against expected production, performance ratio, availability, all faults with duration and resolution, work completed against the preventive schedule, and issues requiring the owner's decision. A report that states the system is operating normally is not a report.

Warranty administration. Module, inverter and battery warranties require somebody to file claims, hold the commissioning records and maintain serial registrations. This obligation should sit explicitly with the O&M provider, because a warranty nobody administers is a warranty that quietly expires. This is one of the most common ways owners lose value they have already paid for, as covered in why solar systems fail early.

Exclusions. Read them carefully. Common exclusions are force majeure, damage from third parties, grid faults, and consequential loss. Some are reasonable. An exclusion for soiling, or for "acts of nature" broad enough to cover dust, is not, in a market where dust is the primary loss mechanism.

Term, termination and handover. What happens at the end of the contract, and what does the incoming provider receive? A complete asset register, maintenance history, monitoring access, warranty documentation and spares should transfer without obstruction. Contracts that make handover difficult are protecting the contractor, not the asset.

O&M priced against what underperformance costs. On most commercial systems a few percentage points of lost annual production exceeds the entire maintenance fee. Sources: Budgeting for Solar PV Plant Operations and Maintenance, NREL and Sandia National Laboratories, 2015; MIMAH analysis.

What Actually Fails, and How Soon You Find Out

Designing the right scope requires knowing what the equipment does over time.

Soiling is the largest routine loss in this region and the fastest to accumulate. It is also entirely recoverable, which is what makes cleaning intervals the highest-return item in most contracts.

Inverters are the components that fail most often. They are the most electrically stressed part of the system, they run hot, and they contain the most electronics. Inverter availability dominates system availability, which is why spares strategy for inverters deserves specific attention rather than being folded into a general clause.

Connections and terminations loosen through thermal cycling. A loose DC termination creates resistance, resistance creates heat, and heat loosens it further. This is why annual thermographic inspection earns its place: it finds the fault while it is still a production loss rather than a fire.

Modules degrade slowly and predictably, at roughly half a percent of output per year in normal service, but they also fail in specific documented modes. International field surveys of photovoltaic module failures catalogue backsheet degradation, junction box failure, cell cracking, hotspots and delamination, and these are found by inspection, not by monitoring, because a single failing module in a large string barely moves the plant-level numbers.

Batteries, where present, are the shortest-lived and most expensive component, and their life depends heavily on operating temperature and depth of discharge. A contract covering a system with storage must specify battery monitoring, temperature logging and a clear position on who funds replacement.

Mounting and structure need periodic inspection for corrosion, fastener condition and clamp position, particularly in coastal environments and after any severe weather.

The detection question matters as much as the failure itself. A total inverter outage on a monitored plant is known within minutes. A single string offline can go unnoticed for months without string-level monitoring. Two dozen modules gradually degrading are invisible without inspection. The monitoring granularity specified in the contract determines which of these you find, and when.

Making Multi-Site O&M Actually Work

Everything above assumes one site. Portfolios behave differently, and the difference is not linear.

Operating a hundred sites is not a hundred times the work of operating one. It is a different activity, and the plants that fail are usually the ones where nobody recognised that. Three disciplines make it tractable.

Standardisation. A portfolio built on a small number of equipment configurations can be supported with one spares holding and one set of trained technicians. A portfolio assembled from whatever won each individual tender cannot be supported at all past a certain size.

Local capability. Sites that are hours from the nearest technical centre are maintained by somebody local or they are not maintained. Training site-level operators to clean arrays, perform visual inspection and identify when to escalate is not a supplementary activity. It is the maintenance model.

Structured reporting. This is the constraint that surprises people. Maintenance reporting that works fine at twenty sites collapses at two hundred, because manual paperwork does not scale and because the value of the data lies in comparison across sites, which is impossible when every report is a differently formatted document. Systematising reporting is what makes a portfolio-wide warranty and maintenance obligation administrable rather than nominal.

MIMAH's own O&M portfolio in Sudan covers 110 solar sites totalling 1.39 MW commissioned in 2025, all remaining under maintenance, and the ASCENT programme adds a contractual one-year warranty plus two years of free maintenance across a far larger station count. The lesson from that work is consistent: the engineering scales easily and the administration does not, unless it is designed to.

The same is true for solar pumping stations, where each site is remote, unattended and critical to a farming season. The design and maintenance considerations are covered in our solar water pumping guide.

Operating a portfolio without standardised reporting? Our maintenance and asset management services cover multi-site O&M, remote monitoring and structured condition reporting. Our project record documents delivery and ongoing maintenance across health, agricultural and industrial sites.

The seven provisions that separate an O&M contract that delivers from one that buys an annual visit. Source: MIMAH engineering and contracting practice.

Frequently Asked Questions

Is an O&M contract necessary for a small commercial system? Some form of maintenance arrangement is, though the scope should be proportionate. Even a modest rooftop system needs cleaning at a defined interval, annual thermal inspection of DC terminations, and monitoring that tells somebody when production falls. What a small system does not need is the full instrumentation and response regime appropriate to a multi-megawatt plant.

What is the difference between availability and performance ratio? Availability measures whether the system was online and capable of producing. Performance ratio measures how much of the available solar energy was actually converted into delivered electricity. A plant can be 100% available and perform badly, which is exactly what happens when it is dirty. Both belong in the contract.

How often should panels be cleaned? It depends entirely on the site's soiling rate, not on a standard interval. Dusty, low-rainfall environments can lose a substantial share of output between cleanings within weeks, while humid coastal sites with regular rainfall need far less. The right approach is to establish the site's actual soiling rate from production data and set the interval from that, then review it seasonally.

Who should hold the monitoring data? The owner. Contracts where the monitoring platform belongs to the contractor create an artificial switching cost and leave the owner unable to verify performance independently. Direct platform access and clear data ownership should be non-negotiable.

What happens if the O&M provider misses the guarantee? Whatever the contract says, which is why the remedy clause matters more than the guarantee number. Service credits, liquidated damages, or compensation for lost energy at a defined rate are the usual mechanisms. If the contract states a guarantee with no consequence for missing it, no guarantee exists.

Should the installer also do the O&M? There are arguments both ways. The installer knows the system and holds the as-built information, which shortens diagnosis. But there is a conflict where the fault originates in the installation itself, and an independent provider is more likely to identify design and workmanship problems honestly. What matters most in either case is that the provider is established enough to still exist in year five and holds the records to administer warranties.

The Asset Only Pays if It Performs

The financial case for solar rests on a single assumption: that the system produces roughly what it was modelled to produce, every year, for twenty-five years. Every payback calculation, every PPA tariff and every business case depends on it.

Nothing in the physics guarantees it. Modules degrade slowly and predictably, but soiling, faults, loose terminations and failed strings are neither slow nor predictable, and none of them announce themselves. The gap between a system that delivers its business case and one that quietly delivers 80% of it is not the equipment. It is whether anybody is measuring, and whether somebody is contractually obliged to act on what the measurement shows.

That is all an O&M contract is: the mechanism that turns an assumption into an obligation. Specified properly, with both guarantees, a defined remedy, real response times, spares that exist and reporting the owner can read, it costs a small fraction of what underperformance costs. Specified as an annual visit, it is a line item that buys nothing.

Have a system whose production you cannot verify, or an O&M proposal you want reviewed before signing? Talk to our engineering team. We will benchmark actual production against what the site should deliver and tell you where the gap is coming from.