23andMe, who does it profit?
Combining data analytics and healthcare raises interesting questions of ethics. In a context where consumer data are an increasingly valuable currency globally and concerns for individual privacy becomes widespread, what is 23andMe’s situation?
Created in 2006, by Anne Wojcicki, 23andMe is a direct-to-consumer company that offers genetic testing for self-curiosity. Upon order, customers receive a kit that consists of a plastic tube and instructions on how to send back a sample of their saliva to one of their labs. It takes 3 to 5 weeks to receive a full DNA analysis report which includes ancestry, inherited traits and health information such as risks of cancer, diabetes, Alzheimer and Parkinson’s diseases, etc…
In order to provide that information, the technology relies on matching the DNA sample received to its database which implies that the larger the database becomes, the more accurate the results will be. It is the key to their business’s success. Otherwise, a negative outcome occurs when for example some percentage of a customer’s ancestry result is classified as “unassigned”, in which case, some of them express their unsatisfaction such as in the following comment left on 23andMe’s blog.
The company is well aware of these kinds of issue and in order to grow and diversify their data base simultaneously, they offer a program called Global Genetics Project for volunteers whose all 4 grandparents are originating from one of the countries listed on their website. Volunteers can participate by donating their saliva and answering a survey so that “23andMe and its contractors [can] access, analyze and store” the sample. In return, participants “receive a complimentary 23andMe saliva collection kit and the option to access the 23andMe Service.”
When it comes to detecting diseases, as stated on their website, “accuracy was determined by comparing results from this test with results from sequencing. Greater than 99% of test results were correct. While unlikely, this test may provide false positive or false negative results.” Another reason why the racial diversification of their database matters is because of how the test performance varies according to ethnic groups:
“Geneticists have studied more people of European descent than other groups, so they have more data on white people. 23andMe knows this, so its reports will include warnings such as that the test results are “most relevant for people of European descent” (for Alzheimer’s) …” (Stat News, Begley).
A good way 23andMe increases their network effect is by how they send updates to their existing customers based on new discoveries made using their DNA sample stored within the company. As 23andMe research technologies improve and as there are able to detect new information, customers are emailed new reports, even years after having purchased the service. It is a great way to engage with customers on the long term and increase the possibility that it will bolster more conversations around the product and trigger word of mouth effect within their customers social circles. Another long term investment to made concerns cloud security to avoid piracy and maintain consumer’s trust and sell more kits.
However, what is the business really is about and who really is the client? For example, 23andMe has a partnership with Lark Health, a startup company that offers diabetes counseling to 23andMe customers. It is easy to imagine other commercial usages of customer’s health information. In fact, the kit’s prices have kept falling since the beginning, from 299$ to 199$ and right now, 99$ for a limited promotional holyday offer. Today, their massive DNA library is the “world largest genetic research database”. Below are two comments made by 23andMe top managements in different interviews that says a lot about the company’s future plans:
Patrick Chung, a 23andMe board member: “The long game here is not to make money selling kits, although the kits are essential to get the base level data” […] “Once you have the data, [the company] does actually become the Google of personalized health care. […] Genetic data on a massive scale is likely to be an extremely valuable commodity to pharmaceutical companies, hospitals, and even governments. This is where the real growth potential is.” (Fast Company, Murphy)
Richard Scheller, 23andMe’s chief scientist: “I thought it was genius actually, that people were paying us to build the database.” (Forbes, Carson & Chaykowski)
Sources
https://www.23andme.com/howitworks
https://www.statnews.com /2017/04/07/what-you-need-to-know-about-23andme-genetic-test
https://www.scientificamerican.com/article/23andme-is-terrifying-but-not-for-the-reasons-the-fda-thinks
https://www.fastcompany.com/3018598/for-99-this-ceo-can-tell-you-what-might-kill-you-inside-23andme-founder-anne-wojcickis-dna-r
https://www.forbes.com/sites/bizcarson/2019/06/06/23andme-dna-test-anne-wojcicki-prevention-plans-drug-development/#322f32c4494d
Super interesting! The ethical concerns you raise reminds me of a case we did in Strategy class called Earth Bank Codes. The company is building a genetics database for animals in the Amazon,but unlike 23andMe compensates those who help build the database. Others like the Human Genome Project make such data publicly available given its great utility as a great public good. I wonder if there are elements from these models on how data is collected and distributed that 23andMe can leverage without jeopardizing their existing business model.
I first heard of this startup when some friends came up saying “I just learned I am 5% of north African descent, isn’t that cool?”. Knowing that my friend is originally from Chile, with a clear Spanish first and last names, I was very skeptical. And I guess I still am. Although nothing prevents current technology advances to enable decoding your entire DNA and figuring out the origin of your ancestors tracing several generations just from analyzing a small sample of saliva, I keep having this feeling that the company may be just making the results up. “Oh you are 80% European and 20% Asian”. That’s an easy guess they could do based on zip codes and other data from your public profiles. And how can you verify that the followed the actual, thoroughly designed, big data-driven and robustly calibrated AI enabled process to really figure out your DNA? Maybe they do that for just, say 10% of the customers, the rest they just do random (0,1). Because how can you audit the results they send you?
Thanks for sharing about 23andMe, Corine! I actually did this saliva test two years ago with the Chinese equivalent of 23andMe. Based on what I just read here about 23andMe, I think both companies have pretty much the same testing methodology. Admittedly when I paid for the test I didn’t expect to get highly accurate results from them because the way the company marketed itself is, I would say, half “scientifically rigorous” and half entertaining/fun. And I’d guess that a lot of consumers would have a similar mindset – maybe just try it out and see if I can get some surprising results? Surprisingly though, I think the results from the first report they sent back to me were 80-90% accurate already – my parents come from different parts of China and their parents are also from different areas and ethnicities. So I’m very impressed to see that the test identified all “elements” of my ancestry correctly with precise numbers. Over the past two years I’ve received minor updates on my results – mostly on potential health risks etc., and I think such updates were due to their ever-increasing database and hence more accurate testing out of a growing sample. So I agree that network effect is critically important for this company to deliver better “product quality”.
I do see Short Apple’s concerns about the accuracy or even the way they do test (wow if they do run random tests irresponsibly as you speculated, I’m afraid the company will face significant backlash and huge legal penalties lol). I actually think your friend’s result of “5% North African” might be reflecting a hidden fact he wasn’t aware of, because the Arabs from North Africa once ruled part of the Spanish territory didn’t they? And so some North Africans did move up to Spain, settled there, and became Spanish afterwards – some of them further migrating to the Americas 🙂
Wow! As a 23andMe user I would be disappointed if the company sold my data to third parties, even if the data could not be traced directly to me. In fact, I don’t recall what the product says about the privacy of a customer’s genetic information (I probably should have looked more carefully). I think genetic data is more sensitive than my searches on google or my activity on Facebook. It is much more personal. If customers are in fact handing their genetic data to the company forever, I do think they should be paying us. Data of that type is very valuable precisely because it is very personal.
Fantastic article so thank you for sharing. Thinking through this the key question in my mind is what corporate responsibility do we have to people who sign away their data in exchange for information. In my head, the response to the question will determines how much data liability the company is amassing with each competed test.
Interesting post, Corine! I agree with the earlier comments… DNA data is about as personal as it gets, with the ability to provide information about your health and identity, so I think there is something concerning about this business. I imagine that the majority of the purchasers of these kits did not pay much attention to the terms of the privacy agreement, perhaps because it is hard to imagine at this point that our DNA data could be used against us. I think 23andMe will need to take incredible care of the sensitive data they have already collected, ensuring it remains secure and anonymized, if they do not want the public to view them as exploitative in the future.
This is a great post! I think this is an excellent example of the necessity of network effects not only for business… but also for science. This is not what we often consider, as most of us has an image of scientist as a lone person spending countless hours on a lone pursuit of a breakthrough in an academic laboratory bended over a microscope. However, here – the more people submit their DNA samples to 23andMe the more accurate their results will be. So, in a sense, network effects matter not only in terms of presence or absence of multi homing but are the core essence of both the business, the product offered, and future scientific gains that might emerge. Good job!
This was a really interesting read! While I had heard about tests like this, I always thought it was mostly something people tried out for fun and because they were curious.
The model of them selling this as a package of it being a way to find out about your ancestors and also do a medical check up left me with a few questions. I was wondering if there are a set of users who actually use this as a way to do a medical check up rather than are more keen on learning about their ancestors?
23andMe is a really interesting company as they’re navigated their value prop for users from being capable of suggesting alternatives to doctors, to connecting families. Their entire product seems to hinge on accesses such data and since uses 23andMe isn’t a necessity, I’m not sure how much users can take issue with the collection of data.