Backblaze data has no possible relevance to consumer drives in a consumer environment. THG "exposed" this fallacy in an article a year or two ago.
Backblaze rents storage and to make money, they buy / create arrays in the cheapest ways possible.
1. They buy the cheapest low budget stuff on the market
2. They literally had the drives mounted in loose racks with the HDs held in place by rubber bands.
3. The used consumer drives in a server environment. This is like using racing slicks to climb roads in the rocky mountains during a blizzard.
Consumer drives are equipped with a feature called "head parking" which returns the drive head to a parked position so that routine vibrations and desk bumps, don't cause the head to crash into the platter. Servers farms are built ... well not Backblaze's .... on heavy concrete foundations that can not transmit vibration. Therefore server class drives do not need have this feature
It would vastly reduce their performance moving the head away from the data as servers are constantly retrieving or writing data. So a mechanical drive may have the same internal construction, with firmware or other variations that make it better for a particular usage.
That "head parking" feature that is oh so valuable in a consumer environment, is actually a detriment in a server environment. HDs are typically rated for 250 - 500k parking cycles and a server drive can go thru that in 3 months ... big surprise why they don't last. Might as well order soup with a fork and then complain it's difficult to eat.
If you want to determine what to use in a home or office, then the relevant data to peruse would be be how many consumer drives used in a consumer HDs are RMAd. That data is readily available. Each set of numbers is the data recorded for drives RMA'd that were between 6 and 12 months of age, followed by the data for the previous 6 months in parenthesis
http://www.hardware.fr/articles/954-6/disques-durs.html
2017/12/09 reporting period
Seagate 0,72% (contre 0,69%)
Toshiba 0,80% (contre 1,15%)
Western 1,04% (contre 1,03%)
HGST 1,13% (contre 0,60%)
2016/05/13 reporting period
HGST 0,60% (contre 0,81%)
Seagate 0,69% (contre 0,60%)
Western 1,00% (contre 0,90%)
Toshiba 1,15% (contre 0,96%)
2015/11/09 reporting period
- Seagate 0,60% (contre 0,68%)
- HGST 0,81% (contre 1,16%)
- Western 0,90% (contre 1,09%)
- Toshiba 0,96% (contre 1,34%)
2015/05/19 reporting period
- Seagate 0,68% (contre 0,69%)
- Western 1,09% (contre 0,93%)
- HGST 1,16% (contre 1,01%)
- Toshiba 1,34% (contre 1,29%)
2014/11/06 reporting period
- Seagate 0,69% (contre 0,86%)
- Western 0,93 (contre 1,13%)
- HGST 1,01% (contre 1,08%)
- Toshiba 1,29% (contre 1,02%)
2014/04/30 reporting period
- Seagate 0,86% (contre 0,95%)
- Toshiba 1,02% (contre 1,54%)
- Hitachi 1,08% (contre 1,16%)
- Western 1,13% (contre 1,19%)
Now all of the above being what it is, I caution against EVER, for any component, relying on brand names for evaluation. Whether it be PSUs, MoBos, GFX cards or anything else, just about every manufacturer has made some great products and some real bombers. You should therefore peruse the data on individual drives and avid the models with high failure rates rather than focusing on brand names.
However, as the above "real market data" shows, the Backblaze study is in no way relevant to your task. Might as well read a study that says "the plastic forks we bought are no good to eat soup".