Personal tracking devices empower individuals with data and promote self improvement. But what happens to all that data? And who owns it?
A great deal is at stake for the companies involved with their customers’ data – reputation, business model and ethics to name a few. And when the conversation about data privacy began on our blog, we became engaged in listening carefully, considering and analyzing options, and moving the ball forward.
The Two-Way Conversation
In October, we announced the addition of the Data Download feature for myZeo, a direct result of feedback from our Zeolots. Slowly and unexpectedly, the post triggered an outpouring of comments from customers concerned about their data privacy.
Naturally, we decided to jump in. After sharing our initial thoughts, Derek and I had the opportunity to speak on the phone with a number of the Zeo users who left their comments. We also received valuable input from a visit to the VRM conference, a group session at the Mass TLC Innovation 2009 unConference, and email conversations with Ron Ploof and C.C. Chapman (who ran a great session at Podcamp Boston #4 called “What Happens to My Data?”). We would like to thank all of you for the feedback and direction that you provided – it was invaluable.
We are still working through many of the issues that were raised in the process (see ‘Other Considerations’ below), but we are listening and we are changing.
Why Share Your Data?
Let’s start at the beginning; here’s what we are doing with the data and why you might consider sharing.
- We will use aggregated, anonymous sleep data to make improvements to our products. By understanding more about how Zeo users sleep (e.g. which steps of the coaching program show the most results, how frequently users upload their data, etc.), we are enabled to develop better, smarter solutions for current and future customers. This is a win-win for Zeo and our customers.
- We will use aggregated, anonymous sleep data for research on sleep in the home. Our research is designed to advance the science of sleep, with a particular emphasis on revealing information that can be useful to consumers. We plan to publish some of this research, or put some of the more interesting learnings on our website, or use it to provide new sleep tips that our customers have found helpful. By contributing your data, you are contributing to the creation of new insights and knowledge of the sleep of consumers, and this knowledge will then be made available to you. This is another win-win scenario.
- We may share aggregated, anonymous sleep data with other researchers who are studying how people sleep in their own homes. For the first time, sleep data is available to consumers in an affordable, easy to use way, in their own homes. This is a big technological advancement and could have large implications for sleep science. Zeo is not a medical device or program, and is not intended to diagnose or treat any sleep disorders. However, Zeo does offer better (and more) information than could be gleaned from sleep surveys alone. By contributing your data, you’re contributing to the advancement of science. This is a win-win-win for the researchers, Zeo, and you as a beneficiary of their research.
- We may provide aggregated, anonymous sleep data to third parties who can enhance their products or services with it. The database described above is clearly a valuable asset for academics. It should also be valuable for commercial entities interested in how people really sleep. We honestly don’t know yet how valuable this data will be – but we are curious about by the possibilities. If things work out well this could clearly be a win for Zeo and the companies that get access to aggregated sleep data. I argue that there is also a benefit for Zeo customers. The more that Zeo is able to monetize aggregated data and develop meaningful partnerships, the more we can invest and develop our consumer offerings.
And this is worth mentioning in the same vein: we will never share identifiable sleep information.
And If You’d Still Rather Not…
We hope that all the reasons above help you to understand the matter more clearly, but if you’d still rather not participate, that’s ok with us too.
The first step we’re taking as a result of this conversation is allowing Zeo users to opt out of this database. Simply email us at data.privacy@myZeo.com with your preference to be removed. We will then remove your data from our aggregated pool – your access to mySleep data and coaching services will not be affected.
In making this decision, we conducted a good amount of research on how other body measurement companies deal with aggregated data and whether they offered an opt-out; most do not, but a few do (e.g. Polar). To be honest, best practices haven’t really been established yet in this relatively new field of consumer body measurement so there isn’t necessarily a right or wrong way to do things – but we have decided as a company that we will gladly allow our customers to opt-out of sharing their aggregated data if they wish to do so.
Other Considerations
There is much more to consider regarding sleep data – we will continue to listen and change.
This comment thread deals with many more issues than just the interest in opting-out of data aggregation. Concern is raised about the need to use the mySleep site at all, preferring a desktop application and no encryption. We need to understand how this affects our business model (we are a start-up after all) and make an investment in the time it would take to develop. For example, does the data become less valuable to us & other third parties if users operate entirely offline? Should we create additional incentives for you to share your data? Would we lose the flexibility to make quick changes (like the addition of the download to excel feature) by moving away from the cloud? Would any of these hurt the value of our company? There’s still more to learn.
Another outcome from this conversation: although our privacy policy contains all of the information mentioned above, you expressed that there is room for us to improve how we convey that information. We are working on an edit that will make it easier on the eyes and will keep you posted when that is available.
Your Feedback is the Key
We plan to continue to work towards satisfying as many of our customers as possible. The data download via CSV was one result of your feedback (it’s your data – take it with you!) and now we have enabled an opt-out from our aggregated database (it’s your data – you don’t have to share it with us), so please keep it coming! There’s plenty to digest here and, even if you’re not one of our customers, please let us know: what do you think?
Thanks Again,
Ben – CTO & Co-Founder
and Derek – Assoc Mgr Social Media
Derek( ZQ: 83) is our social media superstar; chances are you chatted with him on Facebook or read one of his tweets. He is one of the leaders of Zeology and an all-around friendly guy who loves to talk sleep.
{ 2 trackbacks }
{ 27 comments… read them below or add one }
Personally I was a little put off by the idea that I’m stuck uploading online to get full usage out of the zeo. But I decided that I was mostly bothered by the lack of choice. Given the choice I’d probably still choose to use the data online especially if you were to add social type features (friending people so we can compare sleep data).
For your research I’m curious on how you plan to deal with the fact that your userbase is of different demographics than most sleep studies looking at population norms would use. I would never quality for a sleep study for example because I take medications on certain days only that are known to impact sleep. Currently your custom variables don’t allow easily to mark what days I take them so notes are what I guess I’ll be using, but I’m sure no one wants to individually search through people’s notes for things like that.
Also, I wouldn’t qualify for a control group because I have some health issues, hense the medications, but I’d not qualify for a sleep disorder group because I don’t have a sleep disorder.
It seems a large percentage of your user base have some kind of sleep disorder (could be wrong on this) so you can’t really make generalizations about “sleep at home” because the people who choose to use a zeo are a very self selected bunch. Seems any data would be skewed in that direction.
I’m very interested in the idea of research on populations that would normally be excluded from studies. Often studies are so selected that they no longer really represent the true population’s variety.
Hi Lee,
Thanks for your input. We certainly appreciate your willingness to share and we will continue to try to find ways to make Zeo a better experience for all our customers.
On the research side, you bring up some very interesting and valid questions. We are certainly collecting a very noisy database (there are a lot of questions we don’t ask – we don’t want to make using the website a chore), but what it lacks in specifics, we hope to make up for in volume. It remains to be seen whether our sample could truly represent the population, but we definitely plan to ask these questions and explore just how generalizable our findings would be for the whole population. At the very least, any findings that we present would be quite valid for the Zeo population, and that’s of primary importance for finding ways to provide a meaningful and helpful service to our customers.
As we begin to look into these issues and explore the questions around the data that gets collected, we will certainly be sharing it with you all.
Again, thanks for your input and your questions. And keep checking in, there will be more to share…
All,
Many people on this comment thread (and elsewhere) have requested an open version of the SD Card with unencrypted data.
We just released a Data Decoder Library that opens up the data on the SD card (sorry it has taken so long!). This library is for developers (ie. it doesn’t have a pretty UI) and gives unrestricted access to your data without uploading to mySleep.myZeo.com.
We are listening, we are learning, and we are making changes! It’s you data and you can do what you like with it.
You can check out the release and download it here:
http://developers.myzeo.com/data-decoder-library/
I blogged on the topic here as well:
http://blog.myzeo.com/accouncing-the-zeo-data-decoder-library/
And feel free to discuss the library in our forum here:
http://blog.myzeo.com/forum/zeo-sd-card-decoder-library/
oops.
Delivery has failed to these recipients or distribution lists:
data.privacy@myZeo.com
The recipient’s e-mail address was not found in the recipient’s e-mail system. Microsoft Exchange will not try to redeliver this message for you. Please check the e-mail address and try resending this message, or provide the following diagnostic text to your system administrator.
Sorry about that – this is now fixed.
Not a customer, but interested, age 86. I sleep with CPAP mask because of sleep apnea. I take 5 mg Ambien to counter mask discomfort, seem to sleep well but less than eight hours. Would ZEO get me off Ambien?
CV
Hi C W,
There’s an important distinction there. Zeo is not a medical device, and is not intended to diagnose or treat any medical conditions. Always follow your doctor’s advice on that front.
Zeo is designed more to help you discover how the things that you do during the day may be affecting how you sleep. I think the best explanation came from a Zeo user who recently posted on our Facebook wall:
“Although the Zeo cannot do anything about my Sleep Apnea, I find it a great tool in analyzing the type of sleep I am getting and how implementing a certain lifestyle change, such as drinking less coffee or exercising, might help to combat my sleep apnea and improve the quality of my rest.” (http://bit.ly/8y5Sju)
It’s about the power of sleep data to help people achieve a better night’s rest.
I am generally very protective of my privacy and personal data, yet when it comes to this specific data subset, I have no reservations in sharing it. I don’t see how sleep data is in any way damaging to have public, or any threat, or anything embarrassing. I really could care less if my sleep data was 100% public. In fact I really wish you had user forums and/or some type of social aspects where I can compare notes with others. I think the biggest challenge or disappointment I have with using the Zeo and your system is the inability to discuss my experiences and get answers to small questions. Yes, your support e-mail is available, but some of these little things don’t seem like support “issues” plus it would be more useful to get feedback from a group rather than one person.
Hi Carlos – thanks for sharing your thoughts and for the feedback on the forum. We really like the idea and are working through the specifics. We’ll keep you posted as we make more progress on that front.
In retrospect, I rather wish I would have filled out the Zeo information with false information. Bogus name, address, birthdate, etc. That is, keep the profile applicable (same age/gender), but not associated with my identity.
I really don’t mind Zeo having my sleep data, as long as it’s truly anonymous. Not only do I want the researchers to have anonymized data, I want Zeo to have only that as well.
Imagine if someone is injured in a car accident or industrial workplace, and the Zeo logs are subpoenaed from Zeo Inc. That’s a lot of colorful information for an insurer or litigant to use. There’s no reason I need Zeo to keep this, so it can only be negative risk for me.
So — the important question is this: Does the zeo log file contain the serial number of the Zeo device?
Is the uploaded data uniquely associated with a specific Zeo machine?
The reason I ask is that I would like to “de-register” and start anew, with falsified identity. If the serial number is plastered through the data, then that would be a pretty pointless exercise.
Thoughts?
All of those fears presume that:
Someone knows you use a Zeo.
The case is huge enough to warrant this level of effort.
Some judge is going to consider this admissible evidence for some reason.
The last part is the least tenable.
Not too far-fetched — most people at work know I use a Zeo. It’s fun watercooler chat, right?
Any case is huge to insurance.. that’s what they have lawyers on staff for. I don’t think the bar is all that high for saying “here’s some info that might help our case” and demanding it.
I don’t know what Zeo Inc would require from the court to divulge sleep records, notes, etc. Is it specified somewhere? eBay, for example, will divulge any info on any trade or member just “by request” from law enforcement. No court compulsion needed.
“You can see that the defendant often has two drinks before bed, from his own log. On this day, he logged in late at night, used sudo, and we can assume AFTER he had two drinks..”
I’m just saying — I see only downside to having this information associated with my specific identity. It’s not necessary, and it’s possibly reckless.
I’m still looking for a specific answer to whether the Device itself is uniquely identifiable from the uploaded log file.
Some interesting and thought-provoking comments have appeared since I last joined the debate. I have had several valuable interactions with the nice folks at ZEO and they seem to be making good faith efforts to deal with the whole privacy/data sharing issue, but there are still some puzzling aspects. For example, encryption is an active, deliberate process, so arguments regarding the difficulties of decryption just don’t compute. Also, the software for allowing private, at home use already exists — it’s called the ZEO Sleep Coaching software — just provide a standalone version of that, for a nominal added fee.
Before the debate proceeds further, we all need to understand precisely what “data sharing” means. There is little doubt that the accumulated sleep database will be valuable. The question is, will it be shared with anyone and everyone who has a legitimate (sleep researchers) or illegitimate (marketers of sleep products) interest for free, or will it be SOLD? That’s the rub; sharing my data is one thing, but PROFITING FROM IT is something else entirely. As long as customers are forced to upload their sleep data in order to access it, there will always be some doubt as to the use of it. The mere fact that it exists, and will continue to exist for years, makes it vulnerable to misuse. What happens if ZEO is gobbled up by a bigger fish? Or goes belly up? (I don’t think that will happen, but it’s a legitimate question.) What happens to the database? 5 or 10 or 20 years from now, it’s all too easy for that opt-out checkbox to get lost in the shuffle. Lastly, I have yet to hear a good argument for why ZEO (or any similar company) feels that they have an inherent right to own/possess my data, let alone to use it in such an unrestricted manner. I am the only one with an inherent right — all other rights have to be carefully and thoroughly justified, and must stand up to ethical and legal scrutiny.
Jacie,
I agree with your points complete, below are some of the thoughts I have shared with Derek in an email this morning before seeing your blog.
“While not stated in my blog I share some of the privacy concerns others have expressed. I don’t object to the aggregation of sleep data and would probably contribute to that as long as I was in control of what is shared. Your machine is not cheap and the need to replace sensors adds to this expense. If that revenue is not sufficient to be successful you have some business model issues. If harvesting more revenue from the data is necessary to achieve your business objectives you would be far better off selling software that does not require an on-line connection for personal analysis and could also upload data for those willing to share/sell “their” data. It seems to me a good alternative would be to offer “data/dollar” credits toward replacement sensors as an incentive for users to continue to generate and share data that you would then sell to what ever interested parties you have solicited.
There is a win-win available, you guys have the head start, the customer will demand their win, it will be interesting to see if Zeo gets the other win. “
These are good points, Jackie, and I share your concern about the eventual destiny of these extensive sleep records, and behavior journals.
Consider Gracenote, Inc., formerly CDDB (Compact Disc Data Base). The history is a case study in “evolving terms of service”. Users provided their CD information to a public (open) database, which later changed the rules and profited by locking users out of the same data.
The point here is that the data itself, associated with my specific identity, is likely to far outlive the current terms of service, privacy policy, or the even the company itself. It’s given me pause for thought.
For now, I’ve suspended any journal entries and upload the raw sleep data itself. If I can find a way to start anew with forged identity information, I’d go for that.
Accurate information in allows us to give you more relevant tips/ideas in the coaching program. If you decided that you wanted to change the information that you provided at any time, you can always edit Your Profile (at the top right of the coaching site once you’ve logged in). Our database will reflect the current values that are there.
I hope this helps answer your question.
Zeo doesn’t have an inherent right to own your data, however you agreed to them having the data when you agreed to buy and use the Zeo. Like any other organization, people are free to enter into contract and willingly give them the data.
The online app doesn’t just work as a stand-alone app. While not impossible to make it so, there is significant work required and then there will be a significant support burden. That leads to two questions:
1. How many people will refuse to buy the Zeo because of the data issue, and how many of those will actually buy one just because they got a stand-alone app. My guess is very few, so there would be little value to Zeo as a company to do this.
2. What is the value of those users to the company and the community of users versus using all of that time and effort for other gains? Everything is a compromise; time spend making/supporting a stand-alone app is time taken from doing other things that will help the company and the community of users.
It is well understood that more the data collected, the better service, suggestions and advise can be provided by Zeo. Think Amazon rankings and how useful collective intelligence is. However, the customers are purchasing Zeo devices and are not renting them from Zeo, hence need to full access to all features available in the device, as advertised. Having payed the full ASP for the product, the customer own the data that is generated from the device. Encryption at source, as is done today, renders the data useless. To satisfy Zeo’s need for data to better serve customers, and to be open to customers, I suggest:
1. Release a firmware update, that turns off data encryption
2. Clearly warn customers to keep their SD cards in a safe place, if they would like to protect their sleep data
3. Make the services offered on your website so compelling that users voluntarily upload their data to get the best out there
If Zeo still thinks that it cannot let users have access to their own data that they payed for, here are some possible reasons why:
1. Some undisclosed, secret, future intent to use the data in undisclosed ways
2. Unsure if the company can continue to innovate, so lock the users into what exists
3. Show future investors or future potential buyers (of the Company) that they have so many customers on leash!!!
Zeo, please focus on developing an open community of users who work as your evangelists, not by putting them on leash through proprietary techniques, but through constant innovation and unmatched service — It is a win win.
All I need or want is my personal sleep data to never leave my home. That way there is no way it can ever be released for any reason. I just want the data in some form I can load into a spreadsheet and I’ll take care of graphing it myself. I just need a file format and the data to be unencrypted. You wanna charge me more to unlock the data for me? Make me buy an additional piece of software to decrypt it locally? I can’t see that that is a big development or support burden. But since you have one of the few personal sleep monitoring gizmos out there, you could charge for this and I might buy it. Right now in this litigious society I simply cannot take the chance of my data being out of my control.
I’d be happy to give you my feedback on your privacy policy. I am interested in your product, but I do not plan to purchase at this time, because I don’t want to be forced to send my personal data into the cloud. I have an Omron blood pressure monitor that I can connect to my PC. It gives me all sorts of charts and graphs, raw data spreadsheets, average values by day, night, monthly, yearly, etc. And lets me print out charts to take to my doctor. And it is completely off-line. I don’t have to send anything to Omron, ever. I don’t even need an internet connection.
If Zeo did that, then I’d buy it right now.
I understand your intention to possibly provide anonymous, aggregated info to “sleep researchers” as well as your desire to sell data in order to help make your company profitable. However, we have seen that “anonymized” data potentially can be (and in fact has been) reverse-engineered to provide personally identifiable information, as in this case: http://blogs.zdnet.com/BTL/?p=28745
Or what about hackers? Remember when they were able to access all that heavily-encrypted credit card data? http://www.usatoday.com/money/perfi/credit/2009-01-20-heartland-credit-card-security-breach_N.htm
Also, your promise not to ever divulge personally identifiable information would presumably become null and void if your company was ever bought out by another company, or if you went out of business and your assets were liquidated. Remember Enron? Shortly after they went out of business and legal proceedings started, the company’s emails were entered into the public domain. Right now, I can get on the public internet and do a search for “Enron email” and read all the “private” emails that my friend Jennifer sent while she was employed there. She thought her company emails were confidential, and at the time they were, but unforeseen events and circumstances can result in privacy being stripped away.
I just received my New Yorker,saw your ad and looked up your website. What a wonderful product you have created that could be so helpful for so many people.
Unfortunately I won’t be purchasing until you produce a system where the collected personal data is not shared with your company unless the user expressly agrees to this option for research purposes. The data is too personal,too potentially harmful for so many reasons, as expressed by others in this blog.
I, like others on this discussion board, urge you to modify your product so that the unit can be used entirely in the privacy of your own home with no required access to your data by anyone else.
It would seem that some of the many privacy concerns raised here would be addressed immediately by those of us who have purchased the product by editing our profiles. If the company can do more, fine.
There is one point I would like to add to this debate. I’m not just curious about my own personal data, but about how my sleep patterns reflect what I read in the literature, how these patterns are related to my age and health, in short, I intend to use my skills in data analysis to see how consistent my patterns are to the known facts from extant research. Sure, I can confine that effort to my own data, but I’d learn a lot more by sharing methods and results with others in my age group, or in suggesting hypotheses to those in a position to analyse aggregated data. This product opens the possibility of a lay research collective of sorts, assisted by interested experts. So far I’m not finding like minds in these blogs.
Thanks to this company for its transparency on these issues. It’s refreshing…
I’ve suffered with sleep problems most of my life so I was very excited when I learned of your sleep monitor product, the Zeo. After careful research and reading dozens of reviews two major themes emerged; 1) Consumers had major concerns with your privacy policy and 2) A quality issue and the use of your 30-day guarantee.
I’m writing to let you know that it is for these specific reasons I will not be purchasing your product at this time. Therefore I hope you will please take into consideration the following points and suggestions for improving your product and services:
1) If you are indeed truly only concerned with the collection of data for the betterment of your product and services then I see no good reason to tie each identifiable sleep monitor (which can be referenced to a sales invoice, including name and address etc.) with an online account.
Instead, I would have no problem uploading my sleep data (and ONLY the sleep data!) to your web site and even provide you my age, sex and possibly general geographical location. Why is that not good enough for your records?
2) I wish you would reconsider your practice of encrypting data on your memory cards, thereby forcing customers to upload their data on your web site first before taking advantage of the necessary tools to use that data. For one, I find it a little immoral to charge $250-$350 for a data collection device and then require your customers to give away some privacy to interpret that data, the very reason for the purchase in the first place.
Besides, consider the more popular your device becomes the more likely some computer guru will crack the encryption and others will publish open-source applications to replace your online tools. Lastly, for customer’s peace of mind they can use the data no matter what happens you really should offer a standalone desktop application in lieu of the online app. On your blog you posted, “Would we lose the flexibility to make quick changes?” I think not. Just about every software application I’ve purchased over the past ten years has a feature to “check for updates”. Even with a standalone application I would be happy to upload/share my sleep data with your servers. As long as it includes an opt in/out selection and data uploaded is restricted as mentioned above. Many standalone applications have this voluntary feature and I cannot remember one where I’ve disabled it.
3) The one flaw I find in your 30-day return policy is the fact it’s not tied to the date the item was received. So “30-days” is really a misnomer and gives the customer a false sense they actually have thirty days in which to evaluate your product. It’s a little dishonest.
Also, I was told via phone that you would “extend the trial period a few days for the time to ship a replacement part out”. Again, shouldn’t “30-days” really mean I should have an entire thirty-days to evaluate a product? I’ve read many complaints the time it took to issue an RMA, send it back, wait for your receipt/evaluation of device and ship a new one ate up that 30-days despite any padded time for shipment of a replacement. Please consider these facts and update your policy in writing.
I hope you found this information and feedback useful. I’m aware that similar products are currently under development and I’m curious to learn what changes if any your company will make and if these competitors will follow a similar business model. I wish your company much success.
Thanks for the articulate and relevant comments. We are listening – and we hope that as we move forward we will be able to overcome your barriers for purchase.
Best, Ben
Two issues.!) A number of people in this world do not have a broadband connection and need a PC based download and analysis capability (excel is adequate). I was using your web based system, went on a 2 month vacation where I did not have broadband. Consequently I lost 2 months of data and I am reluctant to restart.
My second issue is that I can’t keep the headband on more than 3 nights in a row. If I tighten it up more than I have been I get a headache! ZEO needs to improve the design. Right now I’m using the pod as a high priced alarm clock!
Private does not mean me and you, it just means me. I purchase the Zeo machine to use; I did not trade my privacy for it. If I did, the machine would have been fee and I would have been part of a reimbursed national study.