How I track my life

Tracking quality of life with Reporter.


Unlike many people who I respect, I don’t believe in tracking for tracking’s sake. It’s not one of life’s intrinsic pleasures. Personally, I only have one reason to track: to increase the quality of my life. How? By identifying things that are either correlated or causally-related to increased quality of life and optimizing for those things.

For some background on why I think tracking to optimize quality of life is more important than tracking anything else (including happiness, mood, steps, calories, fuel, how many bananas you ate, etc), read The Death Bed Game. The gist:

You get one death bed point whenever you do something that you believe will still be valuable and meaningful to you when you’re on your death bed. He/she who dies with the most death bed points wins.

I wrote that post 9 months ago, and while I still believe that reverse-engineering quality of life from what you value/regret on your death bed is right place to start, the actual tracking of those moments isn’t very practical. I’ve found it difficult to make the call on whether or not something I’m experiencing now will be valuable and meaningful to me on my death bed. I suspected that I wasn’t being very consistent in my judgments and that the data wouldn’t really be trustworthy in aggregate.

Luckily, Nick Felton and Drew Breunig released an app on Feb 2nd that provided a new way for me to think about tracking: Reporter.

http://reporter-app.com

Reporter is a great app.

I don’t say that lightly. I’ve built apps that were way worse. And I’ve tried so many. SO many. Many of them are okay, most are terrible… a tiny few of them are good. Reporter is great. Here’s why I think so.

First of all, Reporter is all about randomly sampling your day. The beauty of random sampling is that it’s assumed to be incomplete — there’s no completionist anxiety (at least for me) if I miss a day. It may seem silly but completionist anxiety is a real problem for so many apps… including all the ones that quote Seinfield’s “don’t break the chain!” schpeel as a magical revelation. You aren’t encouraged to keep a streak. You also don’t have to remember to track—it can remind you as many times as you want throughout the day (I set mine to remind me about 8 times a day, but most people would probably be fine with less).

Another great thing about Reporter is that you can have it ask you custom questions. Reporter is only as valuable as the questions you set it up to ask you. Assuming you do set them up right (I’ll go through how I’ve set it up in a bit), making it super easy to answer is absolutely critical. It also allows questions to take answers in the form of a number, a multiple choice, yes/no, a location, people, or just a text string, and charts each question’s answers according to their type. Reporter does a great job of auto-completion when you start typing something that you’ve used before, and even suggests the most frequently used answers so you don’t have to type everything — just tap it and move on.

Another great thing is that they set it up to sync with Dropbox — no need to integrate with an API, you have all the data instantly on your computers (it syncs immediately after each report), and it’s accessible in a format that’s super easy to parse by a growing set of available visualization tools. This is infinitely better than storing the data only on your device, only on their servers, or any variation of the two. This solves privacy and accessibility in one swoop, and will work on Android just as well when/if they go that route.

Another great thing is that it automatically captures a lot of ambient data about you without having to do anything: your location, the weather, how many photos you’ve taken since your last checkin, and how loud it is around you. More data, less work. It feels magical.

Everything about the app is designed to make it quick to check in — Nicholas Felton used the app for a year before launching it publicly, and you can tell that he spent a lot of time thinking about how to capture the most data with the least amount of effort.

It’s designed to be used quickly, and often.

They don’t even store the data on their servers. It’s all yours and yours alone—on your phone and in the export. These are good people.

To top it off, it’s also beautiful.

Only downside (if you’re an Androider) is that it’s iPhone only so far.

But seriously, give it a try. </sales pitch>


How I use Reporter

Reporter comes with a set of default questions about your sleep, how much coffee you drank, where you are, who you’re with, etc. Picking the right questions is crucial to getting value out of the app, and is by far the steepest part of the learning curve / barrier to adoption for the app.

I used the default questions for a week or so to get the hang of it, and have been editing, adding, removing questions for the last month or so. Here’s my current list of questions that are working really well for me.

The biggest chunk of questions are about the objective and subjective circumstances at the moment: How much energy do I have? What’s top of mind? How am I feeling? How alert do I feel? What am I doing? Where am I? Who am I with?

I’m interested in these questions because I think they’re all correlated in some way with quality time. I’m not sure exactly how they’re correlated, yet… which is why I’m interested in tracking them and seeing what turns up.

My primary criteria for every question in this set is to know exactly what I am hoping to learn, and how I would act differently if I knew what I was trying to learn.

What I hope to learn: I hope to find patterns of correlation between what I subjectively label as “quality time” and I’m thinking about, how I’m feeling, where I am, who I’m with, how much energy I have, how much alertness I have, etc.

How I plan to change: The goal is to take this information and spend more time around the things that are correlated with quality time and spend less time around things that aren’t correlated. Easy peasy right?

Self-tracking is only useful if it leads to new self-knowledge and—ultimately—new action.

The key question in the list is the first one: “Is this quality time?”

What is quality time? When I get this question, I ask myself if I am experiencing a moment that is in some way connecting me to a person, an interest, or my own subconscious. I need to be experiencing it in a way that feels intrinsically valuable—an end in itself.

When I was trying to find these moments throughout the day manually, it was very difficult to think back and count them up. But in the moment, as they’re being experienced, it’s much easier. This is the magic of Reporter, it samples experience and therefore I’m able to think about moments in the present while never having to carry the burden of trying to capture every moment, or remember them after the fact.

By default, Reporter is designed to track the frequency of things. This quality question, when correlated with the other questions, turns Reporter into an app that tracks the quality of things.

The clincher

The third set of questions were added a few weeks after the first 2 mentioned above. They are the following:

  1. What circumstances are encouraging quality time right now?
  2. What circumstances are hindering quality time right now?

I generally answer the first question when I answer “Yes” to “Is this quality time?” and answer the second question when I answer “No” to it. But not always — sometimes a moment has a mix of encouraging and hindering elements, even when it is tipped more in one direction than the other.

This third set of questions is what makes Reporter actionable. And therefore, what makes the tracking worth doing.
Encouraging circumstances

This is essentially a list of things that for one reason or another catalyse quality time.

Notice that they aren’t all “active” circumstances. Some of them, like “no meetings” and “a day off” are really just open spaces that give room for quality time to happen.

Some are questionable, like “a couple drinks” and “coding” — they most likely lead to different kinds of quality time (with people, and with interests, respectively), but they none-the-less both help in some way.

Now let’s have a look at things that hinder quality time.

Hindering circumstances

Caffeine crash (#2) is pretty interesting, because it is of course directly related to the top “encouraging circumstance”. If I were to rely only on the positive, coffee would appear to be a purely good influence, but now I can weigh both sides more objectively.

Tiredness (#1) is by far the most hindering circumstance. This is also very interesting because it creates an actionable element of “hey, why don’t I get more sleep?” The fear is that I’d be giving up quality time in the evening… but nowhere in the “encouraging circumstances” do I call out “being up late” as particularly encouraging.

Let’s take this tiredness a bit further—as you may remember, Reporter also tracks all kinds of ambient metrics, including time of day. Even though I may not call out being up late as an encouraging circumstance, I can go through the data and see if the hour of the day is particularly correlated with quality time. Here’s what it looks like:

This is really interesting to me for a couple reasons. The first is that quality time peaks for me around the time I get to work, and right after lunch… then dips pretty drastically in the afternoon, and comes back up in the evening.

Actionable items: schedule my most important work for first thing, and right after lunch. The great thing is that I can attempt to make these changes and then come back in a month and see how those changes affect the data.

Here’s another interesting graph:

Tuesdays and Saturdays are the most conducive for quality time at the moment. I’m surprised that the weekend days vary as much as they do, but I’m probably going to wait another month or two before drawing any conclusions from this data.

What things are top of mind most often when I’m having quality time?

  1. Small talk (having a conversation with a small group of people)
  2. Analytics (catch-all name for work stuff)
  3. House of Cards (we did watch a lot of this in the last month)
  4. Chores (proof that I actually do like cleaning)
  5. My new half-brother (who I recently discovered via 23andme.com)
  6. Niko
  7. Relationships (& related conversations with Kellianne)
  8. Investing (something I’ve been researching a bit evenings and weekends)

What am I doing when I am experiencing quality time?

  1. Drinking (ha)
  2. Talking with friends
  3. Talking with coworkers
  4. Cleaning (again!)
  5. Eating lunch
  6. Riding bikes
  7. Talking with Kellianne
  8. Walking

Note that this is not ordered by frequency of occurrence, but rather by how frequency they are experienced as quality time.

Who am I with?

  1. Kellianne (vindicated from the previous list)
  2. Nobody (meaning I’m alone)
  3. Niko
  4. April Schiller (a friend who visited in February)
  5. Nicole (friend)
  6. Coworkers
  7. Bode and Brasa (friends)

I think this list is interesting, but I’m more interested in comparing the numbers over time to see if I can increase the % quality time with the people who I am around the most.

Other interesting data points

Because Reporter also tracks all of that ambient data, I also know that I have the highest ratio of quality time when the forecast is Clear or Mostly Cloudy. I also know that I prefer it to be between 62.3 and 62.6 °F, I tend to have recently taken 5 photos, my battery is at 70%, and I report on average 7 times on my best days. The weather one will be interesting as we go through the seasons… especially if I correlate it only to reports where I’m not at home or at work. I think this information will become more interesting when there are several months of data to aggregate… their signal to noise ratio is a little lower than other data points, so it will take more time to find patterns. Fun!


Conclusions & what’s next

The most interesting and actionable thing I’ve learned is that I may have been thinking too much about the role of behavior-change, and not enough about circumstance-change. There’s a chance that I could increase the quality of my day if I address my hindering circumstance of tiredness (sounds so fancy when I call it that). In the mean time scheduling my most important work for 9am, 2pm, and 10pm might help too. The key thing is to do something differently, with the purpose of improving quality of life during my afternoon slump, and then to come back in a month or two to see if it is working or if I need to try something new again.

Attempting to change a circumstance (tiredness) can have a more significant impact than attempting to change a behavior (bed time).

Behaviors at one part of the day impact circumstances throughout the day, and the set of circumstances then impact later behaviors, feelings, and decisions.

There’s a cascading ecosystem of behaviors at play here that are all interdependent, and Reporter is helping me uncover these connections in a way that no other tracking app has ever done for me before.

If anyone else decides to use Reporter in this way, let me know and I’ll post my scripts that take files in Dropbox and generate second-order stats about stuff in a follow-up post.

Thoughts, feedback, ideas? Tweet me: @buster.