Evan Savage


In this post, I'll tell the story of how I got started with self-tracking and talk briefly about my first experiment.

My Journey Into The Self-Tracking Jungle #

First, A Video For Context #

If you've already seen my talk on panic attacks, feel free to skip to the next section.

Evan Savage - Panic from Gary Wolf on Vimeo.

Seeking Help #

My success in confronting panic began with a simple yet powerful insight:

I don't know how to deal with this, but someone else might.

Once I had this insight, seeing a psychologist was the natural next step.

As a side note, the former point reflects some of the promise of the Quantified Mass. When you have a specific problem, there is a subtle but crucial difference between

What did others try?


What should I try next?

While answering the first question is helpful, I'd argue that answering the second is an order of magnitude more helpful.

Basic Research #

My psychologist recommended the Anxiety and Phobia Workbook. As a survey of known symptoms, studies, treatments, and experiences, it gave me a much broader set of external inputs to draw on. After reading the workbook cover to cover, the natural next step was now to combine these inputs into something actionable.

I identified specific treatments that seemed easy to implement. In retrospect, my initial list was pretty large:

All of these take at most 30 minutes per day, and many are passive habits that rely on small behavior modifications. Small changes in habit are often more effective than large changes, as they are easier to maintain.

My First Experiment #

Building this list led me to another question:

How will I know if my condition is improving?

This is where self-tracking comes in. To answer this question, I needed to know what I was doing and whether it was working. I decided that I would keep a recovery journal, which I divided into four sections.

  1. Weekly Practice Record: this was an overview of my activity. Every day, I would check off each treatment I successfully followed. I also had areas for weekly goals and notes.
  2. Daily Record of Exercise: every day, I would fill in either the duration and type of exercise or a reason for not exercising.
  3. Food Diary: every day, I would fill in my caffeine, sugar, and alcohol consumption. I would also fill whether I took B-complex and C vitamins.
  4. Panic Triggers and Responses: if I experienced a panic attack, I would note the date and time, the severity, what triggered it, what specific symptoms I experienced, and how I dealt with it.

You can view and print the log sheets on Google Docs.

Keeping these logs took no more than five minutes per day. Tracking mechanisms are most effective when they have low overhead, as this lowers the willpower barrier to using them regularly.

A Diversion On Self-Tracking Design #

How can we design systems when we don't know what we're doing?

Although I cribbed the individual sections almost verbatim from the workbook, their specific combination has some curious results.

The broad view of Section 1 is complemented by the deep view of Sections 2-3. In the data visualization world, having multiple levels of abstraction helps the viewer grasp the whole picture without losing their hold on specifics. By looking at the broad view, I knew my overall progress; by looking at the deep views, I could see the areas I needed to focus on.

Section 4 provides the feedback loop. Without this section, I can't answer my earlier question:

How will I know if my condition is improving?

My self-tracking was very goal-directed: I had a specific problem that I wanted to solve. There is another kind of self-tracking, one that I think many people ignore, and that is exploratory self-tracking.

Imagine this same journal without Section 4. None of the treatments are specific to panic, so they could reasonably be followed by anyone. Without the goal of confronting panic, there is greater room for curiosity. You could add more experiments, play with correlations, ask weirder questions like "what happens if I eat a lot of butter?" Without imposing goals, there is no failure or success, and that can be both a curse and a blessing. The curse is that you might not measurably improve yourself. The blessing is that you might not care!

I believe that, much like a reinforcement learning system, the Quantified Self community needs both modes of self-tracking to thrive.