I was one of those people who thought you could simply submit a sample and get back information such as having a connection to a distant group.
That may still be possible, assuming your distant group is sufficiently 'different'. As the global population intermingles more and more, these differences tend to diminish, and matches may mean less than they used to.
To elaborate a bit on what I mentioned above (and without pretending to be either complete or exact), the 'questions' asked in genetic testing boil down to establishing the presence (or absence) of specific stretches of
DNA, a long chain molecule most of which manifests as chromosomes when tightly wound up for storage. Such stretches, sequences of
nucleotides (the molecular building blocks of
DNA), are part of longer and functionally discrete portions of your DNA, referred to as 'genes'. The total of those is called 'genome'.
The stretches selected for tests ('probes') tend to be associated with properties of interest (ethnic/racial origin, physiological traits, susceptibility to disease etc.). The fewer probes used, the more there is an element of interpretation in analyzing the test results, particularly when the question relates to multi-factorial gene effects, and many are. Things get even more complex when you realize that genes and their expression may be influenced by environmental factors and behavior. Finally, it's important to understand that a genome is not so much a blueprint but a recipe, subject to substitutions due to circumstance, etc. In summary, the bottom line in a good genetic analysis is (1) a judicious choice of (2) a sufficient number of (3) known good probes.
Because of the great variation
* possible in the DNA chain molecule constituting genes, even apart from the fact that certain variations or combinations thereof may not have any noticeable effect, these associations aren't necessarily absolute, but are best expressed in percentages (of a full match). That's also why genetic probes often are selected from variable stretches, to maximize finding specific differences. The more probes you test or the more genes you compare, the greater the chance that you are found to be 'different' from others, until you arrive at the—presumably unique—individual: yourself.
*) Note that
human genetic variation is limited to less than about 0.1% of one's DNA. But because the human genome is so huge (~ 5 billion nucleotide base pairs), that still amounts to very large numbers of possible combinations.