Inside Our Lab: How Product Lab Tested Reviews Every Product

Last updated: 2026-09-14

Every review published on Product Lab Tested begins with a purchase order and ends with a data file. Between those two points sit weeks of controlled measurement, blind evaluation, and scoring calibration. This article documents the full protocol — the facilities, instruments, procedures, and policies that produce the numbers behind every Lab Score we publish. We are making the methodology public because trust in a review depends on understanding how it was produced.

We have heard the same complaint from readers for years: review sites test a product for an afternoon, write a verdict, and move on. Our commitment is different. Minimum active testing per product is 40 hours spread across at least two weeks. For categories like robot vacuums and air purifiers, where real-world conditions evolve over time, the testing window extends to six weeks. The goal is to measure what a product does, not what it does on its best day.

The Testing Facility

Our lab occupies 2,400 square feet divided into four purpose-built rooms. Each room controls a different variable because accurate measurement requires isolating the thing you are measuring from everything else.

The climate chamber maintains ambient temperature between 68 and 72 °F (20.0 to 22.2 °C) and relative humidity between 40 and 60 percent RH. A PID-controlled HVAC loop holds temperature to within plus or minus 0.5 °F during a measurement session. Products that generate heat, consume power, or interact with air — laptops, air purifiers, portable chargers — are tested in this room. Temperature drift introduces systematic error into power measurements, thermal imaging, and airflow quantification. Eliminating drift eliminates that error.

The RF-shielded audio room is a 12-by-10-foot enclosure lined with acoustic foam wedges that reduce reflected sound to below the noise floor of our measurement microphones. Background noise inside the room measures 18.4 dBA, verified quarterly with an NTi XL2 analyzer. All audio products — headphones, earbuds, earplugs, speakers, soundbars — are tested here. The RF shielding prevents wireless interference from Wi-Fi access points, Bluetooth devices, and cellular radios outside the room from introducing artifacts into wireless audio measurements.

The light-controlled imaging booth uses a matte-black interior with calibrated D65 daylight panels (6,500 K, CRI 97) and independently dimmable LED banks spanning 50 to 10,000 lux. Webcams, monitors, and displays are tested here. A fixed camera mount ensures identical framing across every product. Ambient light from windows or overhead fluorescents would contaminate display brightness and color accuracy measurements, so the booth excludes all external light.

The general bench room handles mechanical testing, durability cycling, and product disassembly. Bench tools include a digital caliper (Mitutoyo, 0.01 mm resolution), a precision scale (0.01 g), a Shore durometer, and a motorized insertion-removal jig for accelerated wear testing on in-ear devices.

Calibrated Instruments

Measurement quality depends on instrument quality. We maintain a core inventory of calibrated instruments, each traceable to a recognized standard. Annual calibration certificates are on file for every instrument listed below.

Sound and acoustics. The NTi XL2 audio and acoustic analyzer handles SPL measurement, frequency-weighted noise levels (dBA, dBC), and real-time spectral analysis with one-third-octave resolution from 20 Hz to 20 kHz. For in-ear product evaluation, we use a GRAS 43AG ear-and-cheek simulator fitted with a GRAS RA0045 ear canal extension, the standard fixture for insertion-loss measurement per ANSI S12.42. Microphones are GRAS 40AG 1/2-inch free-field units, calibrated to plus or minus 0.1 dB.

Light and display. The Sekonic C-800 spectromaster measures illuminance (lux), correlated color temperature (CCT), and color rendering index (CRI). For display testing, a calibrated X-Rite i1Display Pro Plus colorimeter provides delta-E accuracy to 0.5 across the sRGB and DCI-P3 gamuts.

Thermal and power. A FLIR E96 thermal camera captures surface temperature maps at 640 by 480 resolution with a thermal sensitivity of 0.04 °C (NETD). Laptops under sustained load, portable chargers during fast-charge cycles, and smart-home devices during peak draw are imaged to identify hotspots. A Hioki PW3198 power quality analyzer records voltage, current, power factor, and harmonics for any product connected to mains power. USB power delivery is measured with a Satechi USB-C inline power meter and a Keysight 34461A bench multimeter.

Air quality. The TSI AeroTrak 9306 particle counter samples airborne particulate across six size channels (0.3, 0.5, 1.0, 3.0, 5.0, 10.0 micrometers) at a flow rate of 2.83 liters per minute. It is the primary instrument for air purifier CADR verification and filter degradation measurement. A Testo 400 universal IAQ probe supplements with CO₂ and VOC readings when relevant.

How We Buy Products

Product Lab Tested does not accept review units, loaner samples, pre-production hardware, or any product supplied by a manufacturer or its representatives. Every product tested on this site was purchased at full retail price from a public storefront: Amazon, Best Buy, Walmart, Target, or the manufacturer's own direct-to-consumer website. We pay the same price any consumer would pay on the day of purchase.

This policy exists because review units are not random samples. Manufacturers cherry-pick units for reviewers, sometimes hand-selecting from a quality-assurance pass that filters out the bottom 10 percent of production variance. A review unit may also carry pre-release firmware that differs from what ships to consumers. Testing a manufacturer-supplied unit tells you what the company wants you to see. Testing a retail unit tells you what a buyer actually gets.

We purchase products anonymously. Orders ship to a generic address with no indication that Product Lab Tested is the buyer. We do not contact manufacturers before purchase, during testing, or before publication. A manufacturer sees our review when the public does.

Blind Testing Protocol

Quantitative instrument measurement is inherently objective: a decibel reading is a decibel reading regardless of the brand on the chassis. But every review also includes subjective assessment — comfort, build quality feel, interface intuitiveness — and those evaluations are vulnerable to brand bias. A tester who knows a product costs $400 will unconsciously judge its fit and finish differently than they would a $40 product.

Our blind protocol removes that bias. Before a product enters the subjective evaluation phase, a preparation technician removes or covers all visible branding with matte-black vinyl. The product is assigned a randomized alphanumeric code (e.g., HPN-C7, EBD-A3). Panel testers never see the brand name, the retail price, or the Amazon rating. They receive the de-branded product with a one-line physical description and nothing else.

Each category uses a panel of three to five testers. Testers evaluate independently and submit scores before any group discussion. Only after all individual scores are locked do we reveal brand identities and open deliberation. This sequence prevents anchoring: a tester who submits an 82 cannot revise it to 88 after learning the product is from a prestige brand.

The panel's subjective scores are averaged and combined with the objective instrument data according to the category-specific weighting formula described below.

Score Calculation

Every product receives a Lab Score on a 100-point scale. The score is a weighted composite of five evaluation categories. Category weights vary by product type because the dimensions that matter for an air purifier are not the same dimensions that matter for a mechanical keyboard.

The five categories are:

1. Core performance — the primary function of the product, measured with instruments. For headphones, this is frequency response accuracy and noise cancellation. For air purifiers, this is CADR and particle capture efficiency. Weight: 30 to 40 percent depending on category.

2. Build quality and durability — materials, assembly tolerances, accelerated wear testing results. Weight: 15 to 25 percent.

3. Ergonomics and usability — subjective panel assessment conducted under blind protocol. Comfort, controls, interface clarity, setup experience. Weight: 15 to 25 percent.

4. Feature completeness — does the product deliver on its marketed feature set? Wireless range, battery life, connectivity, app quality. Weight: 10 to 20 percent.

5. Value — performance per dollar, calculated against the full tested set in the category. A product that delivers 90 percent of the top scorer's performance at 40 percent of the price will outscore a premium product with diminishing marginal returns. Weight: 10 to 15 percent.

Raw instrument data feeds directly into categories 1 and 2. Panel scores feed into category 3. Categories 4 and 5 combine instrument verification (e.g., measured battery life vs. claimed) with editorial assessment. The composite formula is arithmetic: each category score (0 to 100) is multiplied by its weight, and the five products are summed. No curve, no rounding until the final integer.

Timeline and Workflow

A typical review cycle runs three to six weeks from purchase to publication. The sequence is fixed:

Week 1: Products arrive and are logged. Serial numbers, firmware versions, and packaging condition are documented. Each product undergoes initial instrument benchmarking — the first data point in every measurement series. De-branding preparation begins for products entering the blind panel.

Weeks 2 through 4: Active testing. The bulk of instrument measurement, durability cycling, and panel evaluation happens during this window. Products that require extended use — nonstick pans, robot vacuums, air purifier filters — run on a daily-use schedule with measurement checkpoints at defined intervals. The minimum active testing investment per product is 40 hours; complex categories (laptops, smart home ecosystems) regularly exceed 60 hours.

Week 5 (if applicable): Data compilation and cross-validation. Instrument data is aggregated, outliers are flagged and re-measured, and panel scores are finalized. Any measurement that deviates more than two standard deviations from the mean triggers a mandatory re-test with a fresh unit purchased at retail.

Final week: Writing, internal review, and publication. The lead test engineer drafts the review with direct access to the full data set. A second engineer reviews the draft against the raw data to verify that every claim in the text is supported by a measurement. If a discrepancy surfaces, publication is held until it is resolved.

Conflict of Interest Policy

Product Lab Tested earns revenue through affiliate commissions. When a reader clicks a retailer link in a review and makes a purchase, we receive a commission from the retailer. This is standard across the review industry, and we disclose it at the top of every review page.

The commission structure is identical across all products in a category. We receive the same percentage whether a reader buys the top-scoring product or the last-place product. This eliminates the financial incentive to rank one product above another. A review that directs readers to the wrong product would also erode the trust that drives long-term readership, which is the actual revenue engine. Short-term commission optimization at the expense of accuracy is a losing strategy, and we do not pursue it.

No manufacturer has a paid advertising relationship with Product Lab Tested. We do not run display ads from product manufacturers on review pages. We do not accept sponsored content, paid placements, or consideration of any kind from companies whose products we test. If a company whose product we reviewed later approaches us about advertising, the answer is no.

Staff members are prohibited from owning equity in, consulting for, or receiving gifts from any company whose products fall within our test categories. This extends to immediate family members. The policy is reviewed annually and acknowledged in writing by every staff member.

Why This Matters

Consumer electronics reviews exist on a spectrum from unboxing videos to laboratory white papers. We occupy the space between: rigorous enough to produce defensible data, accessible enough to inform a purchasing decision. The instruments, protocols, and policies described above are not theater. They are the minimum standard required to produce a number — the Lab Score — that means the same thing every time you see it.

A Lab Score of 92 on a pair of headphones and a Lab Score of 92 on a robot vacuum were produced by the same methodology, applied with the same discipline, using instruments calibrated to the same standards. The weights differ because the products differ, but the integrity of the process does not vary. That consistency is the thing we are actually selling. The reviews are a byproduct of it.

If you have questions about our methodology, our instrument inventory, or the data behind any specific review, contact our editorial team. We publish the process because we want it examined. Scrutiny is the mechanism by which trust is built, and we welcome it.