A figure of merit is a single number that captures how well a device, system, or method performs at its intended job. It exists to make comparisons easier: instead of weighing dozens of separate specs, you distill performance down to one value that lets you rank alternatives side by side. The concept shows up across engineering, science, computing, finance, and nearly any field where people need to choose between competing options.
Why a Single Number Matters
Most real-world systems have many performance characteristics. A solar panel has voltage, current, temperature tolerance, and durability. A sensor has sensitivity, noise level, response speed, and physical size. Evaluating all of those independently is slow and subjective, so engineers and scientists define a figure of merit that rolls the most important factors into one composite value. The result is a shortcut for decision-making: higher (or lower, depending on the metric) means better.
What separates a figure of merit from a plain performance metric is this composite quality. A single measurement like weight or speed is just a data point. A figure of merit deliberately combines multiple data points, often through a formula, to reflect overall suitability for a specific application. That distinction is important because it means a figure of merit is always tied to a purpose. The same device might have a great figure of merit for one application and a poor one for another, depending on which properties the formula emphasizes.
Thermoelectric Materials: ZT
One of the most widely cited figures of merit in materials science is ZT, the dimensionless thermoelectric figure of merit. Thermoelectric materials convert heat differences into electricity (or vice versa), and ZT tells you how efficiently a material can do that. The formula multiplies the square of the material’s voltage response to temperature (the Seebeck coefficient) by its electrical conductivity, then divides by its thermal conductivity, all scaled by temperature.
In practical terms, a high ZT means the material is good at generating voltage from heat, conducts electricity easily, and does not leak heat back through itself. For decades, the best commercial thermoelectric materials hovered around a ZT of 1. Recent computational studies on experimental monolayer materials have predicted ZT values as high as 6.96 at elevated temperatures, though these are theoretical projections rather than commercially available products. The ZT formula is a textbook example of how a figure of merit compresses three competing physical properties into a single number that guides material selection.
Electronics and Sensors
In electronics, the signal-to-noise ratio (SNR) is one of the most fundamental figures of merit. It compares the strength of a desired signal to the background noise in a system. A related metric, noise figure, quantifies how much an electronic component degrades the SNR as a signal passes through it. A perfect, noiseless amplifier would have a noise figure of 0 dB; real components always add some noise, so the number is always above zero.
For sensors like photodetectors or thermal imagers, noise equivalent power (NEP) serves as a key figure of merit. NEP is the amount of incoming signal power needed to produce an output with an SNR of exactly 1, meaning the signal just barely rises above the noise floor. A lower NEP means a more sensitive sensor. Because NEP depends on the physical size of the detector and the measurement bandwidth, engineers often normalize it into a related metric called specific detectivity, which strips out those variables so you can compare detectors of different sizes on equal footing.
Thermal imaging systems use yet another variant: noise equivalent delta temperature, which expresses sensitivity as the smallest temperature change at the source that the system can reliably detect.
Solar Cells: Fill Factor and Efficiency
Solar panel performance relies on several figures of merit, with fill factor and power conversion efficiency being the most common. A solar cell produces a maximum possible voltage (open-circuit voltage) and a maximum possible current (short-circuit current). In an ideal world, the cell would deliver both at the same time. In reality, voltage and current trade off against each other, and the actual peak power output is always less than their simple product.
Fill factor captures this gap. It’s the ratio of the cell’s real maximum power output to the theoretical maximum (open-circuit voltage times short-circuit current). A fill factor of 0.80 means the cell delivers 80% of its theoretical electrical potential. Higher fill factors indicate better internal performance with less energy lost to resistance and recombination. Power conversion efficiency then takes this a step further by comparing the cell’s actual electrical output to the total light energy hitting its surface. Together, these two figures of merit let you compare panels made from entirely different materials and manufacturing processes on a level playing field.
Supercomputing: FLOPS Per Watt
The computing world’s most recognized figure of merit is FLOPS, or floating-point operations per second, which measures raw computational speed. Supercomputers are ranked by Rmax, the highest FLOPS achieved during a standard benchmark test. But raw speed alone misses a critical dimension: power consumption. A machine that performs a billion calculations per second while consuming a small city’s worth of electricity is less impressive than one that matches it on a fraction of the power.
That’s why the Green500 list ranks supercomputers by energy efficiency, measured in gigaflops per watt (GFlops/W). This figure of merit divides computational throughput by power draw, rewarding machines that deliver more useful work per unit of energy. As data centers have grown into significant electricity consumers, this efficiency-focused figure of merit has become as strategically important as raw performance.
Finance and Energy Economics
The concept extends well beyond the lab. In renewable energy finance, the Break-Even Price of Energy (BEPE) functions as a figure of merit for project developers. It calculates the minimum electricity selling price at which a project’s net present value hits zero, accounting for all legal, financial, and operational specifics of a given framework. If the actual market price exceeds the BEPE, the project is economically viable. This single number lets investors compare wind farms, solar installations, and other projects across different countries and regulatory environments without getting lost in the details of each financial structure.
More broadly, any composite metric used to rank alternatives qualifies as a figure of merit. Return on investment, cost per acquisition in marketing, and even academic evaluation indices that combine citation counts with other publication quality indicators all follow the same logic: compress multidimensional performance into one number so decision-makers can compare options quickly and consistently.
How To Interpret Any Figure of Merit
Three things determine whether a figure of merit is useful. First, it must reflect the properties that actually matter for your application. A thermoelectric ZT value is meaningless if you’re choosing a structural material for a bridge. Second, the inputs to the formula need to be measured under comparable conditions. Sensor NEP values measured at different temperatures or bandwidths can’t be compared directly, which is why normalized versions exist. Third, no single figure of merit captures everything. A solar cell with excellent fill factor might still degrade quickly in humid climates, and a supercomputer with outstanding GFlops/W might have poor performance on workloads that differ from the benchmark test.
Figures of merit are decision-making tools, not final verdicts. They compress complexity into something manageable, but the tradeoff is that they always leave something out. The best practice is to use them as a first filter for narrowing options, then examine the underlying individual metrics before making a final choice.