Sign in Subscribe

1. Where it began

I had wanted to do TikTok videos even before I came up with the meta-experiment idea. My friend and I were talking about joining a live TikTok singing challenge. We both have decent voices, but we’re also afraid to sing, which made it a good exposure idea.

Once I got the challenge in mind, I realized it could be the perfect first experiment. I’ve had this belief that not posting videos might be stopping me from growing my audience and making money as a business. That’s an oversimplification, since there are other variables, but this one feels like a major bottleneck.

2. Inspiration & Setup

ChatGPT pointed me to Tasha Doherty's page, who was also inspired by Anne-Laure Le Cunff’s Tiny Experiments. I’ve barely started the book, but I found a wonderful YouTube interview of Anne-Laure, that I watch over and over again. I decided to model my experiments database after Tasha's.

When I started creating this meta-experiment, I asked ChatGPT to draft a general experiment template, which I’ve since expanded and adapted to my context.

Experiment Template

Title:
Short, catchy, curiosity-driven. (e.g., “TikTok Fear Test” or “Can Scafflow Reduce Homework Fights?”)

Hypothesis (What I’m Testing):
1–2 sentences stating the belief or question.
Example: “If I post one TikTok draft daily for 14 days, I’ll reduce friction around singing on camera.”

Why It Matters:
Tie it back to your bigger experiment (“Millionaire by 50”) and/or PGQ system.

  • What survival pattern am I unlearning?
  • What system/identity shift am I testing?

Pact (Action Plan):
“I will [action] for [duration].”

  • Define the output, not the outcome.
  • Keep it tiny, repeatable, trackable.

Constraints / Setup:

  • Time window (when & how long you’ll do it).
  • Tools, environment, or scaffolding (cue stack, fallback version, etc.).
  • Reset protocol (what happens if you miss a day).

Derailer Watch:

  • What derailers are most likely? (perfectionism, low energy, avoidance, etc.)
  • How will I adapt if one shows up?

Data / Tracking:

  • What I’ll log each time (binary: did it or not, friction level 0–5, emotion after, insight).
  • Optional: reflection prompts (like Spin&Dance’s Behind the Dance bonus).

Success Criteria:
Not “did I make money” — but:

  • Did I complete the pact?
  • Did I learn something I can compound?

Results & Reflection (posted after cycle ends):

  • What worked?
  • What failed?
  • What I’ll change or double-down on in the next experiment.

4. TikTok Singing Experiment Specifics

  • I’ll run this experiment until I’ve posted 30 TikTok singing videos.
  • Timeline is flexible: I might post one per day, or sometimes more. So it could take 15–30 days or longer.
  • I’ll start with a baseline talking video so I can see how my metrics change by the end.
  • To keep the data clean, I’ve decided not to post any other videos (on TikTok, IG, or YouTube) during this experiment. Otherwise, extra practice could skew my results.

I chose 30 videos because of Daniel Priestley’s Youtube video, in which he explains how speed is a key ingredient to success and building confidence is the unlock for speed. He explains that confidence isn’t magic, it comes from data. After about 30 samples, you reach your first level of confidence (~65%), because you’ve seen enough outcomes to predict what usually happens next.

For me, each TikTok post is a data point. By the time I’ve sung 30 times publicly, I’ll have a clear read on my patterns: how much fear decreases, how friction changes, and how my audience reacts. That’s when I can expect confidence to kick in and speed to increase.

5. Data Structure

Inside my TikTok Singing Exposure Experiment page, I created a linked database to hold the logs for each of the 30 videos.

  • This lets me track per-video data without cluttering the main experiment design page.
  • I’ll likely do the same for most tiny experiments.
  • For ongoing, long-term experiments (like the 15-month ones), I’ll need a different structure because logging that much detail in a single experiment would be overwhelming.

You can check out my experiments page for any parts of the actual experiment design and the actual experiment data as it unfolds.

Designing the TikTok Singing Exposure Experiment

Testing if posting 30 TikTok singing videos can reduce fear, build confidence, and unlock audience growth without overcomplicating.