Discover more from Ed Zitron's Where's Your Ed At
The Next Steps For The Quantified Self Are Context and Coaching
Every day I go to sleep on a mattress topper that tells me how good my sleep was last night. I know when I fell asleep, I know when I woke up to go pee, I know things like my heartrate, and all of these other wonderful facts and figures. My Apple Watch tells me how many steps I’ve made, my resting heartrate, and many other metrics too. I know on Peloton exactly how often I’ve been on the bike, my average output, my heartrate, and so on. I drop that into TrainingPeaks, which tells me roughly how much I did in totality for a week or a month or a year. I can tell you exactly how much I’ve lifted on Tonal (which I’ve written about at length) in one workout, or in total, and it’s great. My phone can tell me how much time I’ve spent on particular apps and how much they’ve been draining my battery.
And I’m increasingly finding that this information is totally useless.
Okay, let me take a step back. It’s not that it’s totally useless, it’s that we now have a problem where we’ve created mammoth systems to measure and quantify just about anything in our lives, creating multi-billion dollar companies telling us what we’ve done but very rarely prescribing what we should do as a result. Nowhere is this more present than in fitness products, which have become increasingly metric-heavy, spitting out numbers that increase and decrease without much in the way of actual context or information that may make them more useful.
The quantified self movement comes from the idea that you can use devices (wearable or otherwise) to measure things that you’re doing - physical activity, time spent doing something, and so on. It’s very focused on fitness, but it moves into everything from screen time to tracking legal hours now, to the point that it’s fairly easy to know everything you’re doing on a slightly depressing and worrying level.
My sleep is the one place where I’ve felt this the most. I use the Eight Sleep Pod Pro Cover, which I adore as it allows me to be cool while my wife is cold, but it has failed to explain why the last week or two I’ve been extra tired. The answer is potentially in the data - I’ve been sleeping lighter toward 6 or 7am, which may mean that my later bedtimes are getting me less deep sleep. The app did not tell me this, and indeed tells me that I am getting great sleep and, in fact, should be mentally and physically at my peak throughout the rest of the day. The reason that this analysis isn’t great is that it’s not really applying any kind of context to the data, and not really analyzing what’s happening beyond me being asleep for X hours and waking up at Y time.
Peloton has a similar problem. The multi-billion dollar exercise company does absolutely no analysis of your actual rides beyond “you’ve done a certain amount of these, so we’ll recommend more.” The result is that I hired a Peloton coach - yes, they exist - to build a routine for me, and said coach ended up getting me in a state of constant pain and injury…because he was mostly just throwing stuff at the wall and not analyzing anything I was doing. I’ve ended up having to sit and work out what I’m doing each week based on what I know of the Peloton classes I’ve done, and building a routine that factors in TSS (Training Stress Score) and ramps up in a semi-intelligent way.
Tonal does this better than Peloton because it can measure how you’re doing on weights (EG: if you’re struggling to finish a rep, if your form is continually off) and lower the weight, or raise it based on how you’re lifting. As a result, it will continually grow or contract based on your strength, and theoretically mold with you, but I still find that it doesn’t push you far enough with the weights, and had I not done the research, I don’t think I’d have known which classes to do when. Furthermore, it has a “strength score” based on your performance on Upper, Lower and Core exercises, but doesn’t guide you toward what exercises could grow said strength score. However, it does use the context of your workouts to suggest others (things that will pair nicely with them) to compliment them, and programs based on what your goals are.
The reason I love Tonal is that it tells you what to do and how to do it based on your actual performance (both in power and in form), and even then it remains opaque as to what specifically tells Tonal to raise or lower the weights. Nevertheless, it’s the first quantified self experience I’ve found that actually seems to learn from what you’re doing and react to it - personal training at scale - and I feel like anything within the realm of the quantified self needs to move toward being prescriptive rather than obsessing over metrics.
In the case of any fitness products, I am surprised that these products haven’t started doing stuff with all the data they have from user inputs. Peloton is a laughably rudimentary product, and I can’t understand why ~7000 miles of riding on that thing has made it just as dumb as it always was. It doesn’t seem like the most complex thing to do - if my heartrate is running higher on particular intervals than I might usually be, what might have caused that? Should I do something with that information? Or perhaps certain outputs are challenging me - perhaps suggest classes that can improve that particular zone of effort? In the case of just about any sleep app, perhaps the analysis needs to be deeper and educational - “hey, go to sleep earlier, you’re waking up earlier for some reason!” Hell, it’d be nice if Destiny 2 spat out numbers that told me what guns I was good at reliably, but that’s an entirely separate conversation.
It’s something that the software industry has taken to heart to some extent. Sales software tells you when you should follow up with someone based on when they read their emails, and customer service software can analyze calls and emails for negative sentiment or consistent problems. Nevertheless, one of my core problems with Substack is I have no idea what the fuck actually works for people - I have no gauge beyond intuition as to why my most successful posts are what they are, despite the fact that there are 150,000 words to look at and actual numbers. Perhaps there’s no correlation at all, or perhaps there’re subjects that people dig into. I have no idea!
I feel like there’s potentially an entire industry in combining the endless reams of personal data we have with some form of guidance. While we don’t want our computers to become our personal nannies that tell us to brush our teeth and eat food, but there is something to be said for the fact that I have tons of fitness data and tons of sleep data, and no idea what any of it means beyond “I do these things.” I have put thousands of miles into a bike but do not have any real path to further strength beyond “this thing kind of worked?” I know that I am on particular apps for a long time, and they drain my battery, but what do I do with that information? Close them? Get a new phone?
Somehow we’ve got companies that are worth billions of dollars that have got there by analyzing and selling our data without much ability to analyze and understand the data we’re creating ourselves. I really do believe that the next big fitness companies - perhaps Peloton is one of them - are the ones that can tell people exactly what to do versus the freeform jazz of “maybe you’d like this class, I guess.” Furthermore, I think all of these various tools like smartwatches and sleep trackers have to inevitably become significantly more prescriptive with what they do. Your iPhone already tells you when to leave for a meeting, after all.
There are obvious privacy concerns - the more data we give to the computers the more data that’s out there - but it’s not like the data isn’t being created already. The next big data empire won’t be built on monetizing people’s data for sale, but in telling people what to do with their lives.