I realize taste is subjective, however, I trust my taste buds. This is kinda a problem with the system where you have to trust your own judgement as opposed to the default. Using the medium average of a small amount of single flavor recipes does have its downside especially if the 10 people really like their flavoring on the heavy side. So I take a chance on Capella Root Beer with a 10 user default of 7.5% (now down to 7%) and make it 80 mls. I mixed it prior at about 2.5% or 3%, and enjoyed the mix, along with a tasty root beer flavor. Trust me on this one, at 7.5% (even 7 for that matter), this goes beyond root beer flavor into something else hard to explain. Maybe more like a tobacco flavor. It is possible I’m vaping Space Drop and got my label wrong but I don’t think so. Anyway going with your gut instinct is sometimes your best option, especially when Capella never did you wrong at 5% or less.
I agree completely about using percentages on the lower side. People using 25% in their mixes 
skews the median average. The median average is useless if one person is using 25% and another is using 2%. I have made 100’s of recipes and not one contains over 15% in total flavors, let alone one flavor. I always err on the low side and am rarely disappointed.
As for the Root Beer, I have not tried Capella and will have to do so. I use Flavor West and like it very much. I like mine “rooty” and actually use a bit of Madagascar Vanilla which is also rooty. Adding Vanilla Bean Gelato makes a nice RB float.
The figures are there to give you an idea you will always have to dial it in for your own tastes. However a couple of things to note which may or may not be correct and may or may not help you out.
On the percentage in recipes side of things (as I understand it)
- Average mixing quantity is all of the uses of this flavour added together and divided by the number of uses.
- Median is the centre of the usage bell curve (which means that outlying numbers have less effect).
- Maximum usage is generally the usage of the flavour in a stone (as in 100% flavours)
On the Single Flavour side.
- Average mixing quantity is the same as above - all of the uses of this flavour added together and divided by the number of uses.
- Median again is the centre of the usage bell curve (which means that outlying numbers have less effect).
The rest is fairly self-explanatory.
Since individual taste varies so much, taste damage from analogues and length of time vaping all have a bearing on these numbers it is next to impossible to give a truly accurate number that will suit all people.
This is exactly why you see it typed 1000000000 times Single Flavour Test your flavour to your tastes and use those number as a guide only.
When I see something like this, I would not trust it… why? Because my experience is that most CAP flavors are strong enough at lower %. I would start around 4 to 5% and see where it goes. Increase/decrease as needed AND, I would definitely start with a 10ml tester. When it comes to first tries with a flavor, I also shake and vape and although I know a flavor can change a lot by steeping, it gives a good indication of the strength. When you’re dealing with a lot of citrus flavors or strawberries, you know that the flavors will fade but generally speaking, it’s still a good indication.
Right after a shake you can still easily adjust your flavor % if you’re not happy with it.
This is something I really like about the flavor pages on ATF. I always have a look there if I don’t get enough data from ELR. The graph is very helpful and, in this case, the numbers are probably much more appropriate for what you’re after.
Looking around in other places can save you a lot of time, frustration and money when it comes to the lesser popular flavors, flavors that have few or no notes, etc.
A good place to look for additional information is also the flavor company’s website. Usually they have recommendations, but unfortunately, you won’t find that for CAP flavors.
There are other good places to find a little more information, for example the company that you purchased the flavor from. However, those recommendations are sometimes way, way off. As an example below, is the recommendation from one of my trusted suppliers Vapable. They say that they recommend starting off at 15% for all Capella flavors 
Not only that, there are a good few people who try it and leave their comments/ratings which are from time to time surprising…
Like Woftam (and a whole load of other mixers on this forum) said, single flavor tests will provide you with the very best information you can get. It takes some time, some effort and lots of extra bottles (if you’re like me and test 20-30 flavors at a time).
Then there’s the additional topic of flavor ranges that can give you totally different experiences, like you already said, pushing the boundaries can get you a very different flavor profile. While it’s not always a good thing, it can from time to time be very beneficial because you’ll get more flavors for the price of 1. You already noticed tobacco flavors in this root beer, which might be very useful in some recipe (I can’t tell, root beer makes me sick). But there are others, like Pink Guava (Flavorah) for example, that will give a great grapefruit flavor when you’re pushing it.
These are the things you find out when you test the flavors with your own equipment using your own taste buds.
You should check out the DIY Vapor channel on YT and watch how he tests his Flavorah flavors, using a bit of base liquid, adding a few drops, shaking/frothing and dripping. Then adding a bit more and continuing this process to find out what the actual usable range is for a particular flavor.
Obviously it helps a lot if you have an RDA for testing that gives you a similar quality of flavor compared to your daily driver. He also has a video on how to make coils for testing so you don’t need to wick it and constantly replace your wick in this process.
And who are these 10 people? Friends of yours?
There it is right there! Seems to me you’re completely over-thinking this thing.
Go with your own taste buds, not those of 10 others.
True, but outliers can really skew a small sample size. If 100 numbers are used then the outliers have less effect than if 10 are used. As noted, I use it only as a guide and start at much less and work up from there to suit my own taste. Cheers.
Best RootBeer hands down is Liquid Barns!
Good replies all. Since the default percentage started in the calculator, I’ve got in the habit of just going with it because it has been spot on for me except for this one time with the RB. Next time if something doesn’t look right to me, I will pause before I hit the enter key and adjust.
This is amazing.