UPDATE: I no longer recommend testing at AncestryDNA. Currently, there are severe limitations to the tools they offer and your options for third-party tools are now being limited. If you've already tested at AncestryDNA, this post will help you use those results but for faster results, consider a site with more tools and segment data like MyHeritageDNA. Some suggestions in posts may no longer be available for AncestryDNA results. This is too massive and changing too fast for me to update everything at this time.

Is AncestryDNA different? Why are DNA results different from one company versus another? Why are ethnicity results different at every company?

I'm not going to address why results (ethnicity or genealogical) are different in this post. I will say they are different and that doesn't make them wrong. If that's what you're looking for, read this post about why your DNA results are not wrong (even if they are different at different companies).

I am going to talk about why using your DNA results (not ethnicity results) is different at AncestryDNA. This is a topic I've gone on tangents on in other posts so I've realized I just need to write a post I can refer to when appropriate.

Why is AncestryDNA Different?

How is AncestryDNA Different Than Other Autosomal DNA Testing Companies?

So why is AncestryDNA different?

Simple.

Segment data.

If that sounds like too much jargon for you, stay with me!

This is about you—Don't be afraid to learn!

If you want to use DNA to help find your family tree, you need to understand some jargon.

You can actually get away with understanding very little jargon and essentially having an elementary school understanding of those terms (i.e. the jargon can be understood with elementary level words, really!)

I'm not saying having a technical understanding of genetic genealogy (using DNA for family history) won't help you, I'm just saying it's not a requirement.

Genetic genealogy is not as easy as the T.V. commercials make it seem but you can handle this.

Understanding what I mean by "segment data" and how that relates to AncestryDNA being different is important to your success.

You didn't take a DNA test just to give some company money, right?

This is about your family history journey. Empower yourself with some select (and simple) knowledge.

Segment Data

Segment data is information about the chunks of DNA you inherited and share with your matches. That's how we know you and the match are related, you share a chunk of DNA (a segment). You can share many segments—you will share many segments with close relatives.

At AncestryDNA you will likely share many smaller segments with relatives that are even 3rd and possibly 4th cousins.

This "more segments" issue is a quirk of AncestryDNA and I'm not going to discuss it in this post because it's not essential when discussing why AncestryDNA is different.

It's a tidbit of information you can store away for now and learn more about later. Just know, if you go from reading this post to looking at your results at AncestryDNA and the same results at another site, it is normal to see a different number of shared segments.

So I've told you segment data is how AncestryDNA is different and you can infer that the number of segments is provided at AncestryDNA, and that's correct.

What AncestryDNA doesn't provide is the actual shared segment information.

Each segment belongs to a specific location represented as a numerical location on one of the chromosomes.

We know how we are genetically related to someone based on knowing which ancestor we inherited each segment from NOT from who we share in our family tree.

You can be related to someone as represented by the shared ancestor in a family tree and that is not who you inherited the shared DNA from. This is common in southern families, like mine, because we have so many shared ancestors, although they may be pretty distant ancestors.

That is why segment data is important.

Why Take a DNA Test?

Did you take a DNA test so it could tell you about the people you already have in your family tree?

Probably not.

You want to find NEW ancestors, people not already in your family tree.

That means comparing your tree with a match's can't help with the majority of the new ancestors.

Let me say that again another way...

Comparing your tree with a match's tree usually doesn't help you with the reason you took a DNA test.

You might find a new ancestor using a DNA match just as you would by comparing online trees (the DNA just helped you figure out who's tree to look at). In this scenario, the match had their tree a bit farther or you were only a generation away from the discovery, already, and could recognize the potential new ancestor and verify him/her.

This is a perilous method as you're relying on the other person's tree being correct.

Most people who have taken a DNA test are not seasoned genealogists (if you're not either, don't worry, you're in good company). For people who have done very little research into their family tree, you might find several new ancestors. This is just like hopping on Ancestry.com the first time and finding "your" family tree.

Let me tell you a quick story about that...

When I was in college, I had already been doing genealogy for about a decade. One day a friend was helping fix my computer and some how the subject of Ancestry.com came up. I didn't have a paid subscription at the time but we hopped on to look at the free trees. There was his!

That's quite something because we were 20 and his mother was actually listed. It was not a small tree but that was certainly his mother. Except that wasn't his grandmother listed as her mother (his grandmother was still living and he knew her, she may even have been living with his mother).

I'm pretty sure that turned him off genealogy, possibly forever.

He found his ancestor in a tree and that tree could reveal lots of new ancestors. However, he also got lucky that he could see there was a problem and didn't just appropriate all those new ancestors without checking the tree's veracity.

That's how relying on trees, attached to a DNA test or not, is. If you have a DNA match, you share genetic material with that person. You are genetically related (with some exceptions when the quantity of DNA is really small---which is why you're warned not to rely on those distant matches alone, instead you need to use segment data).

Your match's tree might reveal new ancestors to you. You are taking the same risk as using trees not associated with DNA matches, though.

Sharing DNA and a name in a tree does not mean that DNA came from that branch of your family. There are a number of reasons you might not have found YOUR new ancestors with an online tree attached to your DNA match.

• The tree might be incorrect. 
• You might not descend from the shared ancestor at all (either of your trees might incorrectly include that person).
• The ancestors listed for the shared ancestor in the tree might not be correct (this is the most common problem).
• You might share that ancestor but the DNA comes from another shared ancestor.

In this last situation, you still get a new ancestor, you're just missing out on another new ancestor.

More importantly, if you use that shared ancestor to look for a match with the "shared matches," you'll be looking in the wrong place and cause yourself confusion.

There are multiple issues with only using AncestryDNA and "shared matches" (i.e. not using segment data) but my point is, even if you get lucky and get a new ancestor because of a match in a tree, it can lead to other problems using your DNA.

The solution is getting segment data.

How to Use DNA to Further Your Family Tree

With segment data, you determine who you inherited a segment from (that involves some time-consuming work and is not the subject of this post).

Once you know who you inherited that segment from, then, when you find a match that shares that segment with you, they have to share an ancestor of one of two ancestors with you. That means you know which branch in your family tree to look in. If they've done the same work, you'd know which branch in their tree to compare to.

They either share the ancestors from your mother's family whom you inherited that segment from or they share the ancestor from your father's family whom you inherited that segment from.

Remember, each chromosome is actually two halves, one from mom and one from dad. Segment data can't tell you which half the segment is from, that's up to you.

"Chromosome Browser" image of chromosome 2 from FTDNA---these are the segments shared between my mom and her brother. Notice how there is only one chromosome shown, because the segment data can't differentiate between what was inherited from mom versus dad.

This is a chromosome I'm "painting" at DNAPainter.com. Notice how this service shows the two halves. That is the purpose of DNAPainter. I've compared my great-aunt's results to close relatives (both maternal and paternal first cousins on this chromosome, chromosome 7) to identify their large shared segments as coming from the shared grandparents.
Because I know these close relatives belong to the maternal or paternal side (and only to that side), I can tell which segments belong to which half of the chromosome.

Why do shared segments help when trees don't?

You use known relatives to determine who a segment is from (usually a "larger" segment). Then you can use matches who's relationship is unknown but share that segment to find new ancestors. This is not simple but it works.

Using shared segments is similar to working with the "shared matches" tab at AncestryDNA but without segment data, it is common to have a group of people who are all each other's matches but actually don't share a single ancestor.

People who share a segment of DNA share a single ancestor and this is the difference.

Groups like southerners and colonial American descendants will often share several ancestors or possibly such distant ancestors (i.e. too distant to be helpful) that there will be large groups of "shared matches."

Some of those people will share an ancestor, possibly on one segment, but possibly on several distinct segments (i.e. a few people will share one segment, others will share a different segment, each segment came from a different more distant ancestor and therefore will help with a different problem, if you have the segment data, of course).

However, some of the people in that large group of "shared matches" will not share an ancestor. This works like this.

The Problem with Shared Matches

You share ancestor A with person 1, 2, and 5. You share ancestor B with person 1, 3, and 4. Your "shared matches" for person 1 will include matches 2, 3, 4, and 5. Match 2 and 5 have an ancestor in common and matches 3 and 4. However, matches 2 and 3 don't, nor do 2 and 4, 5 and 3, or 5 and 4.

You'd love to find the ONE ancestor you share with all these shared matches, right? Beware!

Checking the shared matches of other members of a shared matches group is an easy way to spot problems.

This is an extremely simple example that you could probably straighten out but in reality, you'd have four or five ancestors (A through E) and groups of two or more matches for a dozen or more matches (so number 1-12 or more). I'm not typing that out because your head would spin.

Color-clustering is a way to deal with this.

How to Use AncestryDNA without Segment Data

With my southern roots, I developed my own technique, the 4 Buckets Technique, to deal with exactly this problem. I had lots of large groups of shared matches for my relatives who were tested.

Many were clear at a glance that the match (i.e. the match I clicked on to see that group of shared matches) was from a particular branch. Some were extremely unclear, though.

There was an additional problem even with the clear groups.

Usually I knew they likely belonged to a broad group. I'll give you a specific example to make this easier to understand.

A major grouping is when the shared matches indicate a match is "probably a Patterson." That means they were my paternal grandfather's family.

However, the Pattersons married the Ledfords, my second great grandparents were a Patterson and a Ledford and second cousins. So "probably a Patterson" could also be a Ledford.

The Pattersons also married the Taylors. No relationship between my third great-grandparents but her brother married his first cousin. So shared matches to Taylors could be Pattersons or vice versa.

The Pattersons also married the Townsends. Ug, I can't even summarize that one easily. His sister married her brother and his brother married their cousin, and then some branches not in my direct line married each other, yeah, BIG groups of shared matches. There are so many options for which branch might be the source of the shared DNA!

Then there are all the times the Ledfords married Taylors or Townsends or some other of our cousins from that side and vice versa.

I've mentioned that geography doesn't help me with DNA. I'm actually thankful when matches are NOT from the county where my Patterson family is from.

If they are, it means I will usually see related families in at least four branches of their family (not necessarily the four grandparent branches but four branches of their displayed tree, which will show between eight and thirty-something possible branches or lineages).

"Probably a Patterson" does not help me find new ancestors using DNA. There is so much inter-marriage in the related families, I usually want to quit.

And just to make it clear, in my direct ancestors, I only have one set of kissing cousins in that fourth of my family tree (that's half of my great-aunt's tree and I usually use her results to narrow the options).

So my pedigree isn't so intertwined, the DNA is intertwined when I compare to matches. DNA is useless without comparing to something.

My point is, you do NOT need to have an extreme case of endogamy or pedigree collapse in your own family to have this issue. It's important to be aware of the implications.

Clustering (I actually call it "bucketing") solves this. It will not prove the relationship, but it gives clear indications of what to try next.

You could think of it like a hypothesis generator. You still need to test the hypothesis, it's not an answer, but you won't have dozens of possibilities that are all equally likely (which is what a list of shared matches gives you).

What About Normal People (non-endogamous)?

You might be thinking you won't have such a hard time if you don't have kissing cousins in your family.

You won't. But you'll still have potential problems.

Remember my hypothetical person 1 with ancestors A and B?

1,2, and 5 descended from A. 1, 3, and 4 descended from B. You descend from A and B but don't know who they are.

Your shared matches with 1 are 2, 3, 4, and 5.

You would be wasting time looking for the shared ancestor between matches 2 and 3 (maybe those are the only two with online trees).

It's OK and a good idea to see if there's a shared ancestor in the trees of 2 and 3 if they have public trees that are quick to check. Without segment data, you can't tell that you should stop, though.

If those are the only two public trees, it's understandable you'd try to work with those two matches.

Once you understand the problem, you at least have a better idea of how much work you should reasonably try and what your next steps should be.

Note that color-clustering should make the two unrelated groups clear. Auto-clustering can be done for AncestryDNA matches and is pretty easy when you're not your own grandpa.

What Next

One of the next steps is emailing all the shared matches in that group to try and get more information (for AncestryDNA, you will actually be using their messaging service since they don't provide segment data or email addresses).

You can read my post about contacting AncestryDNA matches for more ideas.

For additional ideas of what to do next, you can read this post about what to do while you wait for your DNA matches to respond (you can also do these just as part of using your DNA results, you don't have to be waiting on someone to respond).

Hopefully, I've convinced you that segment data is important.

There are tons of people who have taken an AncestryDNA test and a lot of them have public family trees. This makes AncestryDNA appealing but it's also why it can lead you astray.

Learning more about using DNA is necessary. You don't need to identify the shared ancestor for every match you have.

You probably do want to find some new ancestors. You're going to need segment data to take this beyond a certain level.

It's not always fast and easy but there are lots of easy ways to start and progress.

Keep learning and working one step at a time.

---

I highly recommend auto-clustering if you've tested at AncestryDNA. If you've given it a try and using the results takes more effort than you can muster right now, you can try my free program, The Road to DNA Success. Step three in this program is "The Faux Buckets Technique" and simple and spreadsheet-free alternative to my 4 Buckets Technique.

You won't get as amazing results with this program as you would with color-clustering (including the 4 Buckets Technique or an auto-cluster tool) but it is an easy way to get started and doesn't require you learn a lot of new skills to make it work. You can also add auto-clustering to The Faux Buckets Technique to get better results.

Learn more about The Road to DNA Success program, here.