For me, the two most important performance factors are App launching and web browsing. Downloading files can also be included in this because you wouldn't like it if your downloading is taking longer due to your AV. But it's less important. AV slowing down internet speed with or without VPN is a bigger issue.
No one really cares about app installation slowdown. An app taking 10 seconds to install with one product while 12 seconds with another doesn't matter because you are not installing new apps every day. Your already installed apps are not updating every day either.
In app launching and web browsing, looks like almost everyone is very fast in this test except Avast, F-Secure and Trend. App launching slow-down is something I have always faced in Trend and for Avast it doesn't happen always but they often slow down even 7zip, Windows Photos which doesn't happen with any other product. F-Secure slows down the launching of unsigned apps after every single signature update. I have said this before and other users verified it.
When people say that Microsoft Defender is fast on their PC, most of them are not wrong because basic day-to-day tasks like app launching, web browsing, downloading are not impacted by it.
People are also not wrong when they say MD is heavy because it surely impacts performance in some scenarios especially if it's a low config PC. I found a relative of mine using a Dual Core PC with 2gb ram running Windows 10 64 bit and MD was making the PC unusable. I had to turn off many things, uninstall pre-installed apps and install Panda Free to make it light.
Based on my experience, I would say the PC mark score in this test somehow is the most accurate representation of their performance impact on most PCs.
ESET, K7, Panda, Norton these top 4 makes sense to me though I have talked with [USER=99014]@Trident[/USER] regarding this and something he told me is definitely correct which is products like ESET, K7, Panda don't monitor as many things like API calls, do journaling etc. related things as the likes of BD, Kaspersky, Norton, Avast. It's easier to be light when you're doing less and relying highly on signature, heuristics, cloud hash checkups. So here I would give Norton bigger credit to be able to keep things light without compromising too much.
So I would say, a certain part of the test matches with my own experience.
But a performance test like this doesn't give the whole picture. It says browsing is very fast with basically every product but many members on the forum said that they can feel Kaspersky slowing down their browsing. Also, some products use more CPU than others while browsing, launching apps, etc. I know a guy in real life whose laptop's battery life increased after he switched from Kaspersky Free to Bitdefender Free (Later to MD for a different reason). Why? Because all he does on his laptop is browsing and Kaspersky is more resource intensive in browsing. These are some facts that no test shows but could be an important factor for many.
It's not easy to test the performance impact of AV products for a variety of reasons.
