AncestryDNA Surpasses 20 Million

According to Ancestry.com‘s corporate page, they now have more than 20 million tested people in their DNA database.

Open Photo

This is great news!  Every addition to the databases makes it easier to advance our genealogy research.  And while database growth has slowed significantly since FamilyTreeDNA granted law enforcement access to their customers, the market continues to expand.

How does AncestryDNA stack up to the competition?  They lead the pack!  This graph gives you some idea.

At last report, 23andMe had sold more than 12 million kits* (as of May 2020), MyHeritage had more than 4.5 million people in their database (December 2020), and FamilyTreeDNA was thought to have about 1.4 million autosomal DNA users (January 2021).

Of course, those numbers are out of date to varying degrees.  I extrapolated the database sizes based on recent growth rates for a more current comparison.  I estimate that 23andMe has sold about 13.8 million kits, MyHeritage has about 4.8 million in their database, and FamilyTreeDNA has about 1.5 million.

That means AncestryDNA’s database is about as large as the other companies combined!

That’s why I always recommend that genealogists test with the two largest databases—AncestryDNA and 23andMe—first.  From there, they can upload their data file into MyHeritage‘s database for free to get access to additional matches.  (There’s  a small, one-time fee for access to additional tools beyond the list of DNA relatives.)

 

Note

* 23andMe’s “About Us” page says they’ve “sold more than 12M kits”, but their recent presentation to investors said they had 11.3 million genotyped customers.  In other words, their reported value may include outstanding kits that haven’t yet been analyzed.  AncestryDNA and MyHeritage both report how many people are actually in their databases.

32 thoughts on “AncestryDNA Surpasses 20 Million”

  1. I’m curious to know if any testing services break down customers by predominant ethnicity. For example, I suspect Ashkenazi Jews are overly represented on FTDNA given some long standing projects and a push at Jewish genealogy conventions.

    1. That would be useful to know! None of the companies break down their database sizes any metric (country, ethnicity, etc.) FTDNA may have a higher percentage of Ashkenazim, but given how much smaller they are than the other databases, they’re not my first recommendation. They haven’t had the best matches for any of the Jewish cases I’ve worked. And most Jewish surnames are so recent that even yDNA isn’t much help.

      1. Projects are created by individuals. Not always the corporate site. And for different subjects. Surname. Town. State. County. etc. Not every one wishes to join a particular project. Y DNA then haplo group then Str and some have chosen to test/refine those individual groups.

  2. For what it’s worth, I thought I’d add the following reminder. As 23andme’s database grows, our match list does not necessarily grow. It’s been my experience, that because of their cap, as more close cousins test, it displaces those below 25cm on my match list.

  3. I have been on Ancestry for some time. And had my DNA done 5 or more years ago. Little response from any matches that I have connected ..

    1. Depends on their advertising campaigns. Sometimes Ancestry focuses JUST on ethnicity. My matches from those times get their results and then turn off, never looking at Ancestry again. Because a cousin in common here and there has confirmed that is what those people did. More recently family and personal identity have been highlighted in advertising here, so the problem is likely to be more that the person is a beginner. I use an approach similar to Leah’s but not as successful, so I need to make some adjustments.

    2. Yes, that is true. Those who tested on the Ancestry site, loved the commercial for it. Scottish kilt or Viking Sword, etc. Many were only interested in their ethnic breakdown. They will not reply, they are NOT interested in the genealogy site of which DNA can also tell them. How many x number of cousins at different levels, 1st, 2nd, 3rd, etc. Too many are finding out that they are NOT who they think they are. Many long held family deep dark secrets. Some found out that they have 250 brothers or sister, (sperm/egg donor babies).

  4. The thing is, I have access to thousands of matches on Ancestry. On 23andMe, I may have thousands of matches, but they only show me a limited number (1500?). So how does their huge database benefit me — they could have 100 million kits, but I would still see 1500, right? Or am I misunderstanding?

    1. The larger the database, the more likely you are to find matches close enough to advance your research. Consider your top 20 matches at each company. Which are most genealogically relevant?

    2. Remember too that 23andMe is NOT a genealogy based business. X number of years ago, in their company’s statement, that they were a RESEARCH business. They SELL to BIG PHARMA among others, for BIG BUCKS. last year o r the year before, sold to one for 300 million, (of course with customer consent.) But being related to those people. your information could be researched/accessed.

      1. 23andMe is a biomedical research company that also offers some genealogy features. They have always been open about that, and they only do research with the DNA of people who have given explicit informed consent. If you haven’t given consent, your data is not used in the research, period. Not even by proxy of your DNA relatives.

        FamilyTreeDNA, on the other hand, is selling access to their database to law enforcement (including overseas) without getting explicit consent from each user.

  5. Numbers don’t tell the whole story. Ancestry does not provide chromosome-specific breakdowns, nor does Ancestry allow aggregate download of matches.

    1. That’s true, although a chromosome browser isn’t needed for 95% of genealogy questions, and you can use the DNAGedcom Client for aggregate downloads. By contrast, I love the chromosome browser at 23andMe, but they are terrible for trees.

    2. I do like the chromosome browser, of course. My genealogical questions are all at least 5 generations back, looking for parents of 3rd greats at least. I am missing seven 4th great grandparents. But I should give 23andme another download — I import to GDAT (was Genome Mate Pro) — so if I imported frequently, I could keep those matches that get rotated off.

      1. Yes, if you import the matches to GDAT, you won’t lose them. I like the chromosome browser at 23andMe, too, but I find it’s easier to have everything in GMP/GDAT.

  6. “At last report, 23andMe had sold more than 12 million kits* (as of May 2020)”

    “but their recent presentation to investors said they had 11.3 million genotyped customers. In other words, their reported value may include outstanding kits that haven’t yet been analyzed. ”

    If they haven’t been analyzed in so many months, I doubt they ever will be. Perhaps they’re unwanted gifts. I would suggest using the number of genotyped people for your graph so that the numbers are comparable between companies.

    1. Comparing apples to apples would be ideal, but the investor presentation is the first time we’ve ever seen the number genotyped at 23andMe (as opposed to number sold). I know of no way to retroactively retrieve # genotyped to update the past entries in the graph, and I don’t know whether 23andMe will continue to report it.

  7. “About Us” page says they’ve “sold more than 12M kits”, but their recent presentation to investors said they had 11.3 million genotyped customers. In other words, their reported value may include outstanding kits that haven’t yet been analyzed”

    Part of the kit/customer difference may also be due to some customers being genotyped on more than one chip version.

  8. Responding to “, I could keep those matches that get rotated off.”. When I realized I no longer could reach such matches through 23andme, nor use their tools to validate triangulation of such matches, I concluded they were no longer providing much value. For the most part I started deleting these from my database.

  9. Yes the growth is interesting. Though i do not agree with the argument test at the largest two companies. Test at the company which is largest in your region. And in continental europe myHeritage is by far the most dominant company. Uploading the dna is not without error. Because the used snp’s are different, imputation errors can occur especially with matches under 60 cM. Discovered several false possitives. To be not dismissed imediately I am an academically trained bioinformatician with lots of experience in DNA research

    1. The imputation algorithm at MyHeritage does create false positives and can also inflate the size of real IBD segments. If you don’t expect many matches at AncestryDNA or 23andMe, testing directly with MyHeritage is reasonable.

  10. Testing in the right place requires some information. Talk to people on forums about their experiences. For example, my local (Australian) cousins of (East) German origins test mostly with Ancestry as they totally dominate advertising here. But I get far more related matches from people in Germany / Europe from MyHeritage at present.
    Exactly as the forums suggested would be the case.
    Paradoxically, it is Australians who have to push to make a connection back to the emigrant generation who have usually pushed further back in their tree than the Europeans. So I have made most distant links through them. When they venture to add a tree.

    1. Paul Woodbury of Legacy Family Tree hypothesizes that genealogy is more popular in immigrant cultures (like in Australia or the USA) because we want to reconnect with our origins. People who still live in their ancestral areas may never have lost those connections so don’t feel as strong a need to research and document them. It’s an interesting idea!

Leave a Reply

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

This site uses Akismet to reduce spam. Learn how your comment data is processed.