Thanks for that, John. I also think that if we were looking for a metric, a KPI that described a critical success factor, a foundational one might be the number of WikiTree members who have indicated on their profiles that they have taken a DNA test of any kind. As of now, there are 207,729 of those, about one for every 48 of the "DNA Connections" listings.
Of course, in some cases those tests aren't very helpful in indicating a probable genealogical relationship. For instance, of the 12,710 yDNA test-takers that we have from FTDNA, 349 have taken only the 12-marker STR test. We have a total of 11,421 mtDNA test-takers from FTNDA, and of those only 6,026, or 53%, are full-sequence mtDNA tests. We've known for years now that HVR1/HVR2 results by themselves generally aren't useful as a form of positive genealogical evidence, that even with mtDNA full sequencing data an exact match could include thousands of living people and that the common ancestor may have lived 70 or more generations ago.
We also know that we have errors in some patrilineal and matrilineal lines on WikiTree. We can see that when the "DNA Connections" panel for an ancestor shows us member test-takers who indicate they have different haplogroups.
The 10 million DNA Connections listings is a great milestone, but I'd really like to see that number (207,729) of DNA test-takers growing. That's where the comparative data come from. 
On a wish list would be a way to develop an automated way to count things like the resolution of yDNA and mtDNA tests taken. That gets complicated, though. It would likely require additional data fields for WikiTree to be able to differentiate between tests and resolutions, various whole genome sequencings being a good example. We can separately count "Other auDNA" test-takers (there are 2,502 of those, my own WGS test included), "Other yDNA" (1,074 of those, including older tests like those from Sorensen Molecular as well as people who have listed haplogroups reported from autosomal tests like 23andMe and Living DNA), and "Other mtDNA" (1,373 of those; ditto the haplogroups from 23andMe, et al.), but we really can't tell how useful those are just by the numbers. For example, a WGS test that has had the yDNA data uploaded to and analyzed by YFull is factors of magnitude more useful than a circa 2004 26-STR "Y-Clan" test from Oxford Ancestors.
Which I suppose is also a reminder that the names listed under any "DNA Connections" panel are really hints for further research. We regularly see new members mistakenly assume that WikiTree, like the DNA testing companies, are reporting matches. Since WikiTree neither stores nor analyzes DNA data, it can't do that. So "DNA Connections" don't mean there's a DNA match; they indicate that people connected per the genealogical tree have reported taking a DNA test of a type that might be pertinent in researching that relationship. Super valuable, but the 10 million number doesn't imply that all those profiles are, actually, connected by DNA.
