Ancestry’s Thru-Lines vs. 23AndMe’s Family Tree (beta)

23AndMe Family Tree (beta), example redacted

23AndMe Family Tree (beta), example (redacted for privacy)

Ancestry has another winning feature on their site, while 23AndMe is playing catch up

Let me say that I love 23AndMe.com, more than I do Ancestry.com.

I think 23AndMe has the better science currently; and the interpretive algorithms for 23AndMe’s ancestral origins has seemed more accurate than Ancestry’s. Though, both sites have had major updates to their algorithms each year.

23AndMe Adds the Ability to Keep Track of Your Family Ancestors — Hooray!

Earlier this year (2019), 23AndMe announced that they were opening up their new Family Tree feature to “beta” usage. Okay, it’s beta — we can’t expect perfection. But then, almost immediately, they started advertising this Family Tree as a useful feature. As of the end of November, it’s not that great; and on the support forum, there are a lot of people who are excited by the potential, but disappointed in the current execution.

23AndMe has made some improvements already, using feedback from customers. And that’s really great. They changed it so you don’t need a specific date of birth and death for ancestors, and they made it so you can move ancestors around a little bit. Honestly, though, the whole thing isn’t worth much more than tinkering around.

The primary problem for 23AndMe is that they are new to the family tree game. While their genetic analysis and investigative surveys are pretty good, there has never been any built-in method of keeping track of which relatives belong to a particular branch of the family tree. There’s a section to make notes, and a record of conversations you might be lucky to have with other members. But there’s no map.

23AndMe’s auto-generated Family Tree tries to create a blank template for you, the customer, to fill in as you see fit. The colorful diagram branches that it selects for exploration are supposed to represent different groupings of genetic cousins in their database. This is potentially very helpful for tracking down unknown ancestors via reverse engineering.

The problems: Many, many problems exist with the current beta version. The groupings are approximations of where 23AndMe thinks the link may be between your past generations of relatives and you and your mysterious 23AndMe cousins. These groupings are based on shared snippets of genetic data. All that is fine, but there generations tend to be off — missing space for an additional generation or two. Or the group of cousins looks like it’s under a grand-maternal branch, when actually it should be under a grand-paternal branch. There’s no consistency with the left and the right sides as maternal or paternal either. The other major problem is: You have no choice over which genetic cousins in the 23AndMe database will be shown or compared on this auto-generated Family Tree. If you don’t know the origins of these randomly selected people, then it’s kind of useless. You’ll never be able to fill in the blanks. The company claims that the list is based on the names of your strongest/most recent DNA matches. They also say that, if you list a relationship, the other cousin will be notified of the relationship in some way. I haven’t seen any message to me personally, nor any feedback for the family histories that I added. So, I have no idea if that’s true.

Another feature that 23AndMe added in 2019 was the ability to import a FamilySearch family tree to your profile. Okay, that’s interesting. But for that feature, there was no way to limit, conceal or alter the information that had been imported. So, either you accepted the public display of your FamilySearch info, or you just didn’t use it.

Okay, so 23AndMe is trying to start a Family Tree service. They have access to existing info on FamilySearch. They have the ability to do interpretive graphs. So, then, why doesn’t the new FamilySearch import combine with the mapped out family tree? Beta — it’s beta version. Simmer down now, everybody.

The interpretive Family Tree algorithm here has changed significantly, and the way to update was to go into your Profile settings and deny that you wanted to share your Family Tree data. Then go back and say you did, and after a day or so it would be updated. Unfortunately, all of the family members you added previously were removed and set into an awkward holding box, so you could re-add them.

Really, it’s not ready for prime time. The company can do whatever it likes to boost sales and interest, but the Family Tree service is not worth investing a lot of effort in at the moment. Tinker away, but expect to be frustrated with the beta version’s options.

Example of Ancestry ThruLines (blurred for privacy)

Ancestry ThruLines is Better

I didn’t notice that Ancestry had also built a similar service. It is called ThruLines and it has been in effect for the better part of a year.

ThruLines maps out a similar set of relationships between you and your genetic cousins in their database. The visualization is not as cute, but it is more effective.

Ancestry’s ThruLines will show all of your known ancestors on one page for about 6 generations back (to 5th Great Grandparents). If the person is not know in may say “Private” and have a placeholder in that slot. This is the most interesting area of ThruLines. These Private entities are often the missing pieces of the puzzle of family origin. For example, I have a great-great grandmother with a surname that is different and she has no known parentage. ThruLines has a Private placeholder there, and a connection between this Private entity and several families that are about 100 miles away from where my family would have been located at the same time. It’s very confusing, but interesting, and if it is accurate, then the picture that people have created on FamilySearch is inaccurate.

Again, this is potentially huge — very useful for filling in the blanks about the family history.

The problem is human error. When you get back before 1860, there are fewer and fewer verifiable records. And there were fewer people then but they all seemed to have used similar names. So, it’s really difficult to prove whether one person with a similar name is correct, or if it should be a different person, or if the two people are the same person or not. So, if an Ancestry member created bad info (exs. dates and names and family relationships) and bogus entities are sitting in the wrong place, it’s hard to verify and make a correction. Because you have no proof, you just don’t know. In walks ThruLines, and we are given the impression that this is a highly scientific algorithmic matching system based on genetic matches. But the records don’t match. Soooo…, which is the truth?

Ancestry has had Family Tree diagrams for many years, and in that respect it had dusted 23AndMe as a tool for discovering family origins. All Ancestry had to do was add a way to map out relationships between genetic cousins and match the potentially shared entities from each user’s records. Again, 23AndMe is just getting around to throwing a similar system together. But again, Ancestry wins the argument today, because “it has the receipts.” Ancestry has millions of old records, including census, birth and death records, and the burial websites called FindAGrave and BillionGraves. 23AndMe provide none of that info. But Ancestry only provides these records for a monthly price. FamilySearch, in contrast, provides nearly all the same documents for free. So, it’s possible to use 23AndMe to get names, and then switch to FamilySearch to do the reverse-engineering to look for connections to 23AndMe cousins. It is long, laborious work.

23AndMe Family Tree was supposed to cut that work down, and it may actually fill the bill at some point. Not today though. Ancestry ThruLines matches genetic cousins pretty well from what I can see, and provides “Potential family members” to the mix and also notes that some relationships need you to “Evaluate” the hierarchy. There is some clear confusion with the algorithm in this regard.

Note: With all of those positives, customers have seen ThruLines go down for long stretches at a time, with error messages that make it seem as though the customer has been inadequate or at fault in some way. It is confusing and required clarification via the Ancestry support chat.

Conclusion & Bonus Info

Both services are trying now to offer similar features — and they are definitely trying to copy each other’s best ideas — 23AndMe with this Family Tree, Ancestry with the Traits. 23AndMe has such a huge way to go though with the Family Tree — I’m not sure they will ever fully catch up. I hope they do, because I still prefer 23AndMe as a recommendation; though ideally I would tell someone to do both services and to check FamilySearch, too.

The bonus info that you get from each service is different, though. They both allow you to download your raw genomic data. But 23AndMe allows you to also download the Shared DNA Matches, so you can write your own software to mine and interpret the links.

Ancestry, because it has this treasure of ancestry claims, allows you to export a Family Tree to a GEDCOM file. This can be imported elsewhere and lists family members and generational relationships between parents and children.

FamilySearch doesn’t have genetic data yet to do any matches that way, and does not allow for any type of backup or export of your work. Also, anyone can come along and change what you’ve done (which might be good or bad, depending). Can’t complaint too much: Free is free. FamilySearch does have a button that allows you to see your relationship to any other entity if it can find one. This can be really helpful sometimes. What they need is the ability to trace the connection between entities of your choice, rather than restricting it to your root entity and others.

Leave a Reply

Your email address will not be published. Required fields are marked *