Trackers, data and apathy
I love the promise of the quantified self, but let’s face it, it’s a pain in the backside. Whether it’s the cost and choice of the tech, the recording of the data or not knowing how to make use of that data — right now it’s too hard for most of us to unlock the insights that will truly benefit us.
A while back I read an article in the Guardian about wearables called The future of wearable technology is not wearables — it’s analysing the data. It’s mostly from Riann Conradie’s talk at CES 2015 about LifeQ. The piece is by Samuel Gibbs and essentially sets up idea that the wearable or tracker isn’t the point it’s what you do with the data that matters. For me this is indeed the issue, as it is with any data. But analysis is only part of the story. There are a few hurdles that still need getting over before we see wide spread and long lasting value from the quantified self.
For the dedicated quantified selfer, things are pretty good. Recording data about many different activities is easier than ever before, as the data, sensors and hackable devices are abundant. But it is still squarely for the enthusiast or hobbyist, as getting truly useful data to analyse requires multiple devices and inventive approaches to recording and analysing the data. Also it takes dedication, as there’s a lot of manual data entry and a reliance on spreadsheets. The rest of us can count our steps and general fitness activities but beyond the simple nudge to do more we are lost. It’s getting there with simpler platforms and smarter tech but we need to see a few more advances before general adoption and widespread personal empowerment.
For me there a three main hurdles that are holding back the quantified self.
Which devices, platforms, or systems do I choose
This one is a killer. What do you choose to track your data? Is it a bespoke device that tracks steps, activity and sleep, or do you stick to the phone, hoping the battery lasts whilst lugging it around. Different data requires different sensors and sources. As yet there aren’t really any single gadgets that actually track all the behavioural data you may want to record. At best you can invest in a platform or eco-system of devices, and hope that it covers all your needs or is easy to extend. Just how many different discreet gadgets will I need? Even for those of us who love gadgets and recording our data, we still struggle with wearables.
The problem is that you need to actively manage all the devices (which is really about the different sensors) to record the data. They need to be with you and you have to make a conscious effort to ensure they are charged and working. It’s a small thing but an essential one. As holes in the data when you don’t track not only effect the data but also effect your morale. There’s something very demoralising about having walked 5Km only to realise you forgot to track it. It’s feels like you’ve been cheated, duped or in someway swindled. Being able to add it manually helps but isn’t the same, it lacks that sense of reward.
Remembering to track or do is one of the biggest problems. It’s the same with any intentional behaviour that’s out of the norm — you need to find ways to nudge yourself to do it. I always think of Jerry Seinfield’s advice for writers; get a wall planner and colour in every day that you write, when you look at it you will not want to see any blanks. Essentially you are nudging yourself. But once those blanks start adding up, it’s easy to fall into a slump. The small failures can end up crushing you when it comes to holes in your data, and there’s no going back to fix it. We need to make it easier for people not to have those holes. Where possible we must make tracking and recording the data simply part of something we are already habitually doing. It’s no good if you need to explicitly do it, because sometimes we won’t. To achieve this, it will have to fit into our routines and rituals. Once upon a time we all wore a watch, it was ubiquitous and almost always with us, now we all carry a phone. But sensors need proximity so we are looking towards the watch again. The thing is, it’s mostly focused on the easiest stuff. It’s the motion based behaviours, and that’s serviced well enough and for most of us already. We can automatically track our movement, our activity and even our heart rates but how does the data from this work with the other data we may want to record?
The more I have to do, the less likely it is that I’ll keep doing it
As others have pointed out it’s hard work to actively keep up the recording of data. Even tracking weight is a manual activity for most people (unless you’ve bought into a system that has smart scales like Withings). Either way it means recording your weight in a format that can later be combined with other data. Plenty of systems allow for different inputs from other devices, but many want to be the record of choice, the single authority for all your data. When you consider all the types of data you may want to record it starts to get crazy. Check out the trackable options on the Apple Health App, looks pretty comprehensive right?
But as many people have pointed out it misses some pretty major health indicators such as menstrual cycle. How’s that for missing a user need for 50% of your audience. Unlikely to be the record of authority for them. There will always be other data sources, other indicators that you want, but may not fit with this or that system.
For most of it there isn’t an automated option, it relies on people knowing the values and inputting them manually. The nutritional data for Caffeine on the Apple Heath App asks you to input milligrams as a data point, I have no idea how many milligrams of caffeine are in my cortado or my flat white. I’m sure there’s a reference somewhere online, if I can be bothered to find it for each nutritional element I feel is important, but really? It’s this kind of thing that causes the trouble, no matter how good our intentions we struggle to maintain new behaviours. We simply can’t adopt 100 new habits all at once and even if some do become habitual, just how much effort/time are you willing to put in? Imagine sourcing and calculating all that nutritional data for everything you eat. Do you even need to? Lets face it all that recording and active self monitoring stops being fun very quickly, unless you’re Nicholas Felton or have a limitless obsessive side. Calories and food intake is the holy grail of the health/dieting sensor market but it’s tricky to automate. There are some interesting new products, like the GoBe calorie counter and the Vessyl that claim to automatically calculate calories. The GoBe is interesting (if it actually works) as it calculates calories post intake, recording what your body has taken onboard rather than being based on product dietary information and guesstimated intake. But it’s another wearable, another bespoke bit of kit to carry, charge and keep on top of. At least it’s a step in the right direction, until we can get smart toilets that do this automatically (health data from sewage). But the real issue is that it only sees half the picture — the calorie data, none of the context of what you ate or when or why. That additional contextual information is likely to reveal far more about the triggers and reasons for dietary blowouts and spikes. As I said it’s tricky, but it shouldn’t stop us trying.
We need to get to a situation where we collect all this data automatically and passively, that way it will be easier to achieve much wider adoption. When I don’t have to think about data input it’s way more likely to get tracked. I don’t want to have to tell something manually that this is a ‘workout’ versus me sprinting for the bus. But at the same time I want that sprint for the bus recorded. I don’t want to see the aggregation of my calories for the day, unless I can see when and what was consumed and in what circumstances. I’m never going to have the discipline to ensure I manually capture that, I need the tech to step in. Ultimately it’s about more sophisticated sensors and nuanced data, and we are getting there. By using a mix of many types of data, the relationships between our behaviours and activities, and the relevant indicators we can gain better insight. I need to understand the things that change me in order to change the behaviours that cause them and this needs a diversity of data.
Seeing the data isn’t the same as knowing how to respond to it
This one is the hardest, as it’s where we need the most help from others. Let’s say we’re happy with our devices, habitualised our data tracking and input, then what? How do we get value beyond the nudge to do more exercise and eat less of the bad stuff. Or is that enough? If we want more, we need analysis and insight to help us understand how the changes we make impact us. We also need a little help in choosing which changes will be the most affective and realistic.
Lets start with the analysis part. Thankfully many of the platforms and services create reports and charts to help you analyse the data. Some utilise the social aspect and play on simple competitive tendencies to encourage more or better activity. Others use your data combined with everyone else’s to understand the patterns and trends in their system, some of which is shared back to you to help contextualise your personal data. But often these are fixed views that don’t allow for data exploration or adding more data to them — they tend to be from inside the one system only.
I use run keeper to track my exercise. It’s a useful App and service, it’s on my phone so fits with how I like to workout (running with headphones and music). It’s a very deliberate and intentional form of tracking and coaching activity. I tell it the context (running) and hit start when I set off and stop when I’m done. It works well for measuring that distinct activity and helps by coaching me through my run. I tried using it for other things like cycling but as I cycle for exercise, fun and transport it didn’t work so well. If I chose to record all bike rides (when I remembered) the data was really skewed and the averages and aggregates tended to be depressing. I either needed to pick and choose when to record, or find some way to slice the data later to make it rewarding. It’s not that I didn’t want to record all the activity, rather that the reports for analysis the system supplied, lacked the nuance for me to ask the questions I wanted. Prepackaged generic views of data only ever get you so far —at best they give you well known metrics useful for the simple nudging around basic goals; run more, improve time etc. They are descriptive — how fast am I, how many times did I run last month, how far have I run? But often the answers we want from the analysis are hints to how we can improve. To answer those kinds of questions you need nuanced data, and often a lot more contextual data. That’s when you can start exploring a hypothesis and often bust a few personal myths. Only then does it become possible to experiment and test ideas to improve. To do that you need to be able to analyse data from different sources and explore the associations in that data. To analyse what other environmental or lifestyle activities are influencing you. That’s why the Apple Health App is so verbose, but it doesn’t help me with the analysis or diagnostics, it doesn’t help me spot where I can potentially intervene and make a change.
Intervention requires hard work and analysis, and some kind of insight into the factors that may be influencing me, as well as ideas on how to challenge them. If I don’t have any ideas, where can I get help identifying this critical things that I can change? If I can’t experiment and try out ideas to improve, I may end up just looking at a bunch of metrics that tell me that I’m failing. I’m going to need proactive advice and I’m going to need those nudges. I think that’s where the network and social thinking comes into play. It’s simply no good being a quantified selfer in a vacuum. Yes you will manage to quantify yourself but to make change you need more. Be that in the form of peer encouragement, help, benchmarking or simply seeing what worked for others. This shared and open personal data platform will be key. It needs to also be on my terms not the technologies, as this data is created by and should belong to me. It needs to be a fair exchange if I share this with a larger entity and I need assurances on use and real value in return. There’s some interesting development in this area and personal data rights. Another take on this is the sharing of health data, with startups like Ubiome and the ‘for the greater good’ aspects of the Apple Research Kit. Interesting times for health, privacy, naval gazing and self improvement.
Over this year I’m going to experiment a lot more with ‘quantifying myself’. I’m going to explore ways of recording data, analysing it and experimenting. Hopefully I’ll be able to get a better sense about the core questions of QS like:
Do I even want all things tracked all the time, just how much do I want to be fully quantified?
What sort of impact can I make on me, being that I know I struggle to keep this sort of thing going?
And of course, explore the gadgets, tech and data analysis that is deeply embedded in the contemporary version of this idea. After all this is fun and not the time and motions studies of mid century industry.
First idea to test will be:
How much influence does the music I run to actually have on my performance? With the natural follow up, what else effects it?
Stay tuned as will post follow up, here.
