How we test laptops

From unboxing to publishing, here's how we perform laptop reviews at Engadget.

Devindra Hardawar/Engadget

Engadget has been reviewing laptops for two decades, and while the definition of what a portable PC is has changed considerably since, our obsession with testing their limits and serving up informative buying advice remains the same: to help you figure out what the best laptop is for you. Be it a hybrid tablet like Microsoft's Surface machines, a rotating 2-in-1 convertible like HP's Spectre x360s or a plain old clamshell notebook, our review process follows similar beats. How does it look and feel? How fast is it? Whether it’s a Windows device, a MacBook or a Chromebook, we aim to answer the most important question in our laptop reviews: Is it actually worth your hard-earned cash?

The review process starts the moment we receive a laptop and start unpacking it from its cardboard cocoon. Though we don’t always mention packaging, we do take note of how well companies protect their hardware during transport, along with how potentially wasteful some material can be. These days, PC makers are more conscious of providing recyclable packing solutions, but we still need to double-check their eco-friendly claims.

A notebook can tell us plenty even before it's turned on. When unboxing a review unit, we immediately note the quality of its construction and its overall weight. A flimsy display hinge and noticeable case flex are instant red flags. We've handled enough machines to tell when an ultraportable starts to feel less "ultra" light and more “ultra” cheap. (We sometimes also use postal scales for accurate weight measurements.)

We also consider all of the included accessories. Does it use a standard USB-C power adapter, or does it rely on an outdated barrel connector (or worse, an entirely proprietary plug)? Additional hardware like a stylus and dongles are worth considering, especially if the device only has one or two USB-C ports.

While some laptops can’t be modified to add more RAM and storage, we note when they can, since that’s a boon for repairability and the machine’s lifespan.

We start our initial testing by seeing how long a machine takes to boot, and note any odd sights or sounds. Are the fans spinning up loudly? Is there a neon LED light show all of a sudden? These aren’t necessarily bad things (especially for gaming laptops), but we’re usually on alert for potential anomalies.

Once we've successfully booted the machine, we’ll load up our favorite apps, which often includes Slack, Spotify, Discord, Chrome and other web browsers. Before the benchmarking begins, it's important to use a computer as if it were our daily driver. How's the web browsing and YouTube experience? What's it like juggling the apps we typically use? This, after all, is the way the vast majority of people will experience their PCs.

To get some hard data, we rely on a variety of benchmarks to push systems to their limits. Across PCs and Macs, our current test suite includes Geekbench 6, Cinebench (both R23 to compare against legacy systems and the new 2024 test) and 3DMark Wildlife Extreme (which thankfully works across iOS, MacOS, Windows and Android). For Windows systems, we also run PCMark 10 and 3DMark Speedway.

We go a bit further for gaming PCs by running 3DMark's Port Royal Ray Tracing test, as well as Cyberpunk 2077's built-in benchmark (which is also a wonderful way to get real-world ray tracing scores). NVIDIA's Frameview app helps us capture average fps figures in Halo Infinite, as well as other games we're playing. We typically benchmark in 1080p, 1440p and 4K (if available) with the highest graphics settings available.

To weigh media encoding capabilities, we also transcode a 4K clip into 1080p using Handbrake on PCs and Macs. By default that gives us a solid understanding of CPU encoding speeds, but we also measure GPU encoding across NVIDIA, AMD, Intel and Apple graphics hardware. The ATTO benchmark helps us determine the average and top disk speeds by transferring large files, which could also have a huge impact on media production.

While pushing laptops to their limit, we also monitor how warm they feel around the keyboard and the bottom. It’s rare to see a machine that will actually burn your legs these days, but we can never be too safe. We also track CPU and GPU temperatures while running 3DMark’s benchmarks.

We observe displays across their entire range of brightness both indoors in typical office lighting, as well as outdoors in shade and direct sunlight. Overall brightness is key, especially when it comes to using the device in sunlight, but we also look closely at color uniformity and any potential display anomalies. We pay even closer attention to high-end display technology, like MiniLED and OLED, to determine the benefits they offer over LED screens.

To ensure our display experience mirrors typical users, we watch videos, browse the web and work in productivity apps. We judge screen reflectivity across a variety of environments, as well as how smoothly they scroll through content with refresh rates beyond 60Hz (if available). For panels with HDR, or high dynamic range, we keep a close eye on the peak brightness and dark scene details during videos.

As for gaming laptops, we stress their support for higher refresh rates while benchmarking games like Halo Infinite and Cyberpunk 2077. If available, we also look at HDR performance within games, which often helps to make you feel more immersed within virtual worlds.

Laptop speakers have come a long way over the last few years, with premium machines like the MacBook Pro packing in six separate speakers and force-canceling woofers for a bit of low end. We listen to music and videos across low, average and high volumes to determine how richly detailed they sound, as well as the point where audio starts to sound distorted. Realistically, we don't expect laptop speakers to sound like a pair of bookshelf units, but things sure are getting close in large notebooks.

There's no excuse for a bad keyboard or trackpad in 2024. We typically type reviews on our test devices to get a feel for key depth, responsiveness and overall comfort. Larger key caps and wider layouts generally lead to a more luxurious typing experience, while smaller notebooks often have limited space for full-sized keyboards. We take particular note of what companies have to sacrifice (like the cramped design of the Razer Blade 14's keys) to accommodate space constraints.

For gaming laptops, we note how comfortable it feels to play with the typical "WASD" control setup in shooters, as well as how easy it is to reach other essential keys. A rare few machines also come with slim mechanical keyboards, which we test while gaming and typing to make sure they actually perform better than caplet keys.

We determine the fluidity and accuracy of trackpads while we work and keep an eye out for common issues: Does the touchpad have any areas where it doesn't feel responsive? Is it stiff to click (especially while right clicking)? Is there enough room to execute multi-finger gestures without frustration?

A lack of ports can almost ruin an ultraportable. We consider the ports PC makers choose to include (or remove) in their laptops and determine if those are acceptable choices for their target markets. A light ultraportable can get away with just having two USB-C ports (especially if it has a separate power port, like the newer MacBook Air models), but that’s unforgivable for a workhorse machine like the XPS 13 Plus.

In general, we like to see multiple USB-C connections with Thunderbolt support, perhaps a legacy USB Type A port if there’s room, as well as a 3.5mm headphone jack. For larger machines, we like to see an HDMI port, Ethernet and maybe even an SD slot, if possible.

We test real-world laptop battery life by counting down how long it lasts during a typical workday (or several) while dealing with web browsing, video watching and productivity apps. For standardized results, we run the PCMark 10 "Modern Office" benchmark on Windows machines, and we test Macs by looping a video until they're entirely drained. We also gauge how quickly laptops can charge up completely, especially if a PC maker touts fast-charging as a feature.

  • We run multiple video calls to judge webcam and microphone performance.

  • While fan noise is less of an issue these days, we still keep an ear out for how loud laptops can get under load.

  • If a manufacturer offers any special hardware capabilities, like the rear LEDs on some Zephyrus G14 models, we try to see if they’re worthwhile.