Most “I tried X for 30 days” experiments fall apart under a simple question: how would you know if it didn’t work? If you can’t answer that, you didn’t run an experiment — you ran a habit and narrated the result afterward.
A self-experiment you can actually trust comes down to four things.
1. A baseline
You cannot measure a change without a “before”. Before you change anything, spend a few days recording the outcome you care about while living normally. That baseline is the control you compare against. Skip it and every later number is floating, anchored to nothing.
2. One change at a time
If you cut caffeine, start a new sleep routine, and begin exercising in the same week, a better night’s sleep tells you nothing about which change caused it. Confounding is the quiet killer of self-experiments. Hold everything else as steady as you can and move one variable.
3. A measurable outcome
“I felt better” is not data. Pick something you can put a number on every day — a 1–10 rating, hours slept, a count, a yes/no. Consistency matters more than precision: the same rough measure taken the same way each day beats an exact measure taken haphazardly.
4. Honest accounting of noise
Some days are just bad. You travel, you get sick, you sleep badly for reasons unrelated to your experiment. A trustworthy result accounts for that variability instead of pretending it away — which is why a difference of “half a point on average” might be real signal, or might be nothing, and only the spread of your data can tell you which.
This is the loop HypoMe is built around: baseline, change, a daily check-in that takes seconds, and a plain-language verdict at the end that is honest about how confident the numbers let you be. Not every experiment will give you a clear answer. The ones that do are worth far more than another month of guessing.