The death of research as we know it #qs2012

I’m sure I’m not the only one who is quite skeptical of research, particularly medical, psychological or economic research. Thankfully, those will be dead sooner than we expect.  

These researches usually go like this “Top researchers at the University of [top-university] found that by [action] 37% of people got [reaction]”. Once you explore you see more information about how they've got to that conclusion. Usually is a survey and they asked 300 people in some city in the US and so on and so forth. The skeptical flag on my head goes up instantaneously: 300 people only? In a single region? Using a self-reported data? To which I’ll immediately dismiss the survey finding as unproven (unless of course I want to make a case and I need to point to something even though I don’t believe on it). 

Small Samples 

The reason researches use small samples, surveys and limited geographic scope is a simple one: money! Research cost money. Lots of money. The more limited the scope, the cheaper it gets. The more “automated” the cheaper it gets, hence the reason they prefer survey to interviews, and interviews to real quantified data. That’s why they survey 300 people instead of 30,000. That's why they prefer to survey top 10 metropolitan cities in the US, instead of the top 1,000 metropolitan cities around the world. 

Quantified Self: n=1 

Well, maybe it doesn’t matter what researchers find what’s the optimum number of steps you should take every day to lose 5 lbs or to lower your blood pressure. You can find out all by yourself. Get a fitbit or BodyBug, start tracking some food intake with LoseIt or MyFitnessPal, and take your weight and blood pressure every day with Withings, and in a couple of months you’ll have strong data about your own body -- even better if you can do over a year or two. That’s unquestionably good data for you, because it’s your data.  

This is what quantified self is about. About you tracking, measuring, collecting and analyzing data about yourself and learning from that data.  Of course, some types of data are much harder to collect and analyze and to actually drive learning and conclusions, and even if you have conclusions about that data it still doesn’t mean is actionable, but it’s knowledge nonetheless. 

Like your financial advisor might disclose to you: Past performance is not a guarantee of future results. So knowing how you did on the past might not guarantee that using the exact same steps will get you to the same place. But what if someone that had a very similar body type, gender, age, race and lifestyle than you also did the same tracking, and she was a year ahead of you. Could you use that information in a way to turn around a problem you are having? Assuming she had that problem and got to a great place? 

n=1,000,000,000 

This is the power of a billion people quantifying themselves. Imagine that we had 1 billion people collecting hundreds of data points on a daily basis about their physical activity, nutrition, sleep, lifestyle choices, stress, mood, location, sex, spending, etc. Imagine all that data gets collected into a single database with trillion of data points and anyone can query that data (anonimized of course). Now, instead of sample-based data we have population data. Instead of a handful of data points for the exact demographic and psychographic profile you have, you have thousands of data points. Now you can look at different paths and different outcomes for those individuals even before you choose your own path (btw, PatientsLikeMe is a company doing that for cancer and other 'big' diseases) 

That's the future of research. It won't be a one-size fits all. It won't be overarching conclusions about good and bad food, good and bad ways to exercise, good or bad ways to save money for retirement, it will be advice 100% personalized to you, based on your data and the data of a billion people. 

Pipe dream? Well, every credit card purchase you make today is logged at Visa, MasterCard, Amex, etc. There are billions of people using credit cards every day. Every phone call you make also gets logged. Not what you said, but who you called and when.More than 2 billion people carry a cellphone on their pocket. Even without GPS, cell-tower triangulation has a pretty good approximation of where you are, and if you are moving, that gets logged too. All saved on a database somewhere. Right now you can't do anything with that data, but the more people quantifying themselves (n=1), the more large companies will see the interest on giving people their data so they can do whatever they want with it. 

Some day, over the next five to ten years, hundreds of millions of people will be proactively sharing their data with services and tools to help improve their lives. Their health, their commute time, their finances and much more. The future looks bright and I won't need to pull my hair out when I read another bullshit headline of some medical research (mostly because I won't have any hair left).