It's not the the free stuff isn't good.... In the November test, Avast Free caught 99.9% of the "widespread and prevalent malware discovered in the last 4 weeks (the AV-TEST reference set)". That means that 11 viruses got thru. Once cleaned over 1200 infections out of a Defender / Essentials protected machine back in 2016 and it was back 3 months later with 670 more. Have had machines come in with all sorts of free and paid for stuff but they wouldn't be here, if they weren't infected. Ya can't fix stoopid and even the best get surprised on ocassion.
The free trials of Kaspersky or BD get most of them but once infected usually have to remove a few manually. But even in most recent test, Defender caught only 98.9% of 0-day malware attacks, inclusive of web and e-mail threats (**Real-World Testing**)", And for December it gained a 100% score But like a medicine that helps 99 people and kills just one ... that's a pretty good performance record... unless you happen to be that one.
I usually clean them with a 30 day free trial of Kaspersky or BitDefender Suites and leave it up to the user to pay for the subscription ... usually about $5 - $7 a seat. Personally, I have too many seats to have to bother with manual upgrades and manual scan runs common with many freebies. We even gave up on BitDefender whan they switched to a model where eachs eat had to be downlaoded and managed from each box. If it comes back, dealt with that once, if you get infected again you are on your own. Have never had a box infected that I was responsible for administering in 26 years.
As for the testing, any test site that doesn't use **real world** testing isn't worth looking at They have a collection of visual in the wild, they install an OS, thy install the AV, they expose it to viruses. Just as important is false positives:
Windows Defender "False detections of legitimate software as malware during a system scan - November = 11 / December = 16) Industry average is 4 not 0
"False blockages of certain actions carried out whilst installing and using legitimate software = 1 and zero here
Industry average is 0 on both counts.
And performance impacts:
"Slower installation of frequently-used applications" November = 42% (26%) December = 54% (35%)" Industry averages in ( ) Just because it was a frequently used app doesn't mean it was downloaded from a reputable place I'd rather it was fully scanned during install than not at all just because it's white listed as a "frequently used app"
That being said, what you use should depend on the value of what's being protected. If it's a gaming box, not really anything there that can't be replaced and Defender is more than enough. If that's the only copy of 1,000s of family photos or 26 years of business records, might want to give up $7 a year. If using Defender, then if something gets by, ya can always download one of the free trials and hopefully get rid of it. Get hit again, well try the other one this time.
But for my money, If it saves me 3 minutes a year, it's cheaper to pay the site license.