← back to blog

To improve your life, first measure it

Or: What I Learned From 3 Years of Personal Metrics

2024-01-02 by Nick Bradford

I zealously tracked every aspect of my life from 2020-2022, then took 2023 off to see what would happen.

Such data! Wow

I learned a few things. But first:

Why did I measure myself?

In January of 2020, I had a minor existential crisis upon realizing that I had been nominally writing a novel for an entire decade, yet had made negligible progress since graduating high school. (Thankfully, the rest of 2020 was entirely crisis-free for myself and everyone else in the world…)

I realized nearly all of the ~400 pages I wrote came from 2012-2013, when I had set a strict rule that I was not allowed to go to sleep unless I wrote 250 words (~1 page). And let me tell you: there is no pleasure known to mortals superior to watching a graph go up.

Novel word count, 2010-2014. The reader is actively discouraged from speculating about the actual quality of the words written by me as a teenager. I’m on my third and final rewrite of Spellcraft as a Service, which you can read here.

If I could gamify my productivity once, I could do it again. And if something is worth doing, it’s worth overdoing. For reference, I had previously inventoried all of my belongings, movies watched, books read, etc., so this kind of thing was fairly in character.

I accumulated several dozen metrics, normalized to a 100-point scale, and simply entered the data manually in Google Sheets every day. I didn’t want to try being as in-depth as Felix, though I do definitely recommend looking at his site. Some metrics were stickier than others, but the most important were:

  • Hobby time: goal of 2 hours/day

  • Sleep duration: goal of 8 hours

  • Wakeup time: goal of 7:30am (bonus points for waking up earlier)

  • Sleep quality: goal of 100% (using a WHOOP tracker)

  • Alcohol: goal of zero drinks/day (“bad” = 3 drinks)

  • Caffeine: goal of zero (“bad” = 400mg)

  • Exercise: goal of 3x/week

  • Personal Morale: self-rated on 1-10 scale at lunchtime and dinnertime

  • Productivity: self-rated on 1-10 scale at lunchtime and dinnertime

  • Calories + macros: goal of 160g protein, 2500 kcal

  • Social media: goal of 10 minutes (“bad” = 30 minutes)

  • Monthly Budget adherence: % over budget

Subset: Raw inputs on the left, short-term metrics in the middle, monthly metrics on the right.

The thick gray line is the monthly moving average of my “Discipline Gauge” (all others are weekly):

Discipline Gauge, 2020-2022.

Disclaimer: Before we delve into my supposed learnings, I want to stress that these are observations about myself, but will NOT necessarily generalize well to other people, and are NOT prescriptive. The only thing I’m prescribing is to take at least 3 years to obsessively collect metrics about yourself so you can find what works for you. ;)

On Productivity

My biggest surprise was how much variance there was in my day to day productivity.

I no longer think of “productivity” as simply increasing my efficiency (work accomplished per unit time), but the combination of the effort I put in and how effective my efforts are.

Productivity = Energy x Tools x Skills

Skills are the foundation, but are gained slowly over time and thus stable day to day.

Tools are great (I think GPT-4 has made me >2x as productive by itself), but new transformative tools are rare, and I find most “productivity” tools deeply suspect.

Energy levels, on the other hand, can vary massively from one day to another.

Given that this blog post is about daily metrics tracking, you may find it unsurprising that my learnings are about how to maintain high energy levels.

TLDR: have a strong goal and stay healthy. I estimate I lost ~10% of my productivity in 2018 (before I started tracking) to lower back pain from bad posture. Yes, very sad. Anyway, how do we get high energy?

Be excited about your work

This factor is so strong I’m tempted to call it “energy/happiness duality”. On days I truly dreaded work, I accomplished literally nothing, and passed the time by checking emails and brooding. On my most enthusiastic days, my energy levels were similar to low-morale days on which I consumed 400-600mg of caffeine (one Red Bull is 80mg).

A secondary reason: a large amount of my output does not come from sitting down “working”, but during downtime (in line for lunch, in the shower, etc), and being excited about something increases the chance of getting my subconscious to do the hard parts.

The real hack I found was to get much better at questioning *why* I wasn’t excited about something. Often the cause was solvable; usually, it was a case of me not deeply understanding the task at hand or believing that it would accomplish the higher-level goal.

An important corollary from this is that when I’m not excited about something, “just buckle down and push through it” is usually the *opposite* of the right thing to do, because it will likely result in me doing something improperly or failing to solve the higher-level goal.

The other big reason I might not be excited is: because I’m too tired! When everything “logically” seems good but I’m not excited to work, I have tried both descending into a state of existential horror and taking a nap, and found the latter more efficacious.

Rest, obviously

I’ve always hated sleep, to the point I used to actively dislike the feeling of being well-rested because it meant I could’ve spent more time awake doing things. I believed you couldn’t say you were working hard unless you were exhausted.

It may surprise you to hear this, but I was in fact wrong. Finally getting enough sleep had such a large effect it was clearly visible in my performance review grades at work. All those studies saying sleep deprivation severely impairs cognition are true, and the scary part is that after you “adjust to less sleep” you just won’t notice that you’re under-performing!

I found one particular night’s sleep to matter much less than my weekly average.

I need an average of 8.25 hours of sleep a night to be fully rested. Below 7.5 hours the effects are noticeable, and below 6.5 hours I’d call my performance “massively degraded”. Relatedly, it took me up to a week of oversleeping to recover from sleep deprivation, but an isolated 6-hour night here and there didn’t have much effect.

I’ve found rest nearly as important on the micro-scale. Trying to muscle through feelings of exhaustion during the day hardly ever worked, but taking a 5-minute break nearly always did.

Diet matters way more than I thought

The most surprising energy factor to me was how “I am what I eat.” Warning: This seems like the area with highest variance from person to person, so your mileage may vary.

Significantly lowering my carb intake (though not to the point of ketosis) made my energy levels incredibly stable and high throughout the day - so high, in fact, that I would often have great difficulty falling asleep! It turns out that carbs help ensure tryptophan is metabolized into serotonin, which is an important part of getting drowsy/relaxed/sleepy. As a result, I now have adjusted to having very few carbs through the day, but a bunch with dinner, and that’s been working well.

I tried intermittent fasting a few times for >1 month at a time, but have never noticed much of a difference either way.

Alcohol was somewhat detrimental, but surprisingly less than I originally guessed.

All in all, I found “eating healthy” overall led to me feeling much better, but that it takes several days to “kick in”/recover after a particularly gluttonous/glutinous (hah) weekend.

Finally, a note on weight: by the end of 2020 that I had lost ~25 pounds in six months (~1lb/week). My secret? Caloric deficit; I ate ~30% less with little change to my exercise routine. I don’t necessarily endorse this method for the same reason it’s hard to endorse calorie restriction to slow aging: you’re hungry all the time and and think only about food instead of productive things.

Weight (lbs) over time.

Exercise is basically a cheat code

I was slightly more energized on days I exercised, and if I went more than 4 days without exercise, my productivity would significantly drop off. But forget that, exercise also allows you to live longer while eating more? And you get most of the benefits from only 1.5-2 hours a week? Sign me up. Plus, people will occasionally give you words of encouragement like “Hey, did you start working out? That’s good, you really looked awful before.”

Sadly, consistency is key, and the chart of the main lifts I prioritized (squats and deadlifts) tell an unfortunate cyclical tale (the top two lines):

Exercise logs.

Don’t do drugs (actually just one drug in particular)

Ah, the most popular drug on the planet, consumed daily by 80% of the world population, and the obvious answer to our energy-boosting quest: caffeine.

Unfortunately, it has been hard to come to any conclusion other than that caffeine gives me energy today by borrowing from tomorrow. This makes it an indispensable tool for bursts lasting up to three weeks, but I find by the end of a cycle I’m consuming >600mg just to function, and my sleep is all manner of screwed up.

In the break-glass caffeine ingestion contingency, I find matcha to be a source far superior to coffee, with significantly reduced jitteriness/anxiety and no crash (due to the L-theanine). It’s so effective, I once took matcha powder with me on a trip to Tulum for late nights out. (I wish I were joking.)

I’m all-or-nothing

If I do something, it’s with 110% effort or with 0% effort. I’ve always been this way.

When I first started collecting metrics, I thought it would be easiest to tackle one area of my life at a time: first hobby time, then exercise, then diet, etc. This didn’t work.

Instead, I changed my mentality to “I will live a disciplined life that fulfills its full potential”, and everything got much easier.

I think of this as creating harmony with your underlying goals: exercising more and drinking less may be different metrics, but because the underlying goal is to be healthy, it creates less cognitive dissonance to simply do both at once.

Positive goals solve your negative goals

I entered the experiment wasting a lot of time on random social media stuff. It wasn’t one of the metrics I initially tracked, but around six months in I decided to add it.

My big surprise: my usage was already down to almost nothing! I had managed to break the vice grip of social media dopamine-hit dependency without even trying. The cause: I was filling all my spare moments with hobby time to get to my 2-hour daily goal.

I found this extended to many other areas. When I said to myself, “I’m not going to drink tonight because I set this rule”, low chance of success. When I said to myself, “I’m not going to drink tonight because oh my goodness I have so much to do tomorrow morning I have so many goals there’s no way I’ll accomplish them hungover”, much better.

One weird trick: let’s think step by step

A hilarious and well-known fact about LLMs is that you can vastly improve their performance by telling them “let’s think step by step” (“take a deep breath and work on this problem step-by-step” works even better), and I’m proud to admit that it works flawlessly on me too.

On Metrics

So: did I finish the novel? Absolutely not! Justifying my procrastination by saying that reading would make me a better writer, I read 50 books in 2020. That’s what I deserve for setting my metric to “hobby time” instead of “writing time”, which leads me to my lessons learned:

The map is not the territory

It is very hard to escape Goodhart’s Law: when a measure becomes a target, it ceases to be a good measure. Metrics are proxies for what you really want. And because we become what we pretend to be, setting the wrong metric can push you to optimize very hard for the wrong thing.

I found myself constantly gaming my own system. If I’m penalized for ingesting caffeine, perhaps I’ll choose methods that don’t list the exact content, so I can justify estimates at the low end of plausibility. If I’m rewarded for all workouts equally, perhaps I’ll do a two-mile run instead of going to the gym.

You can’t have two top priorities

Collecting too many metrics (and setting too many goals) can dilute your attention and cloud your true priorities.

I often think about Galad from the series The Wheel of Time: he set a strict moral code to prioritize his various duties and allegiances, and thus was able to act with great decisiveness.

It’s the destination, not the journey

Metrics should be collected only to help you accomplish a goal, not for their own sake. You should not obsess so much over your process that you forget the content. And collecting metrics isn’t free, for you pay a price in attention, your most valuable resource.

Most productivity habits have a short half-life.

With my metrics, I hit a natural wall in mid-2022 when I switched from a normal job to exploring startup ideas, which made my original first-class goal of increased hobby time totally useless. (Who needs hobbies when you get to work every day on what you think is most interesting? Also, at a startup there is no time for anything other than working maniacally.)

I kept collecting the other metrics, but realized they had stopped affecting my behavior. Things were stable. I had internalized the learnings, and was left with only the overhead.

To be honest, it was a huge relief when I stopped.

In Conclusion

A year without metrics, and I’m happy to report the habits I ingrained have held steady.

For 2024, I’m pausing or simplifying many of the other systems I maintained. I’m streamlining my budget (I don’t know why I thought I needed a separate category for haircuts), and halting maintenance on my personal Zettelkasten (hundreds of pages of personal notes in Google Docs) because knowledge is being organized better by LLMs. Who knows, at this rate perhaps I’ll stop maintaining that list of all my possessions.

Think about your goals, stay healthy, have a happy New Year.