BTT #001 · Experiment brief
Preparing the experimentCan I Replace
This Subscription?
Can an AI-built website help people find useful subscription alternatives—and attract 1,000 unique visitors in 30 days without paid ads?
The hypothesis
People paying for subscriptions may find value in a clear comparison of alternatives, including the tradeoffs of switching. We’ll test the problem and the audience before expanding the product.
The experiment
- Target: 1,000 unique visitors.
- Window: 30 days from the public launch.
- Paid acquisition: $0 planned.
- Channels to test: search, relevant communities, X, directories, and useful shareable content.
What we’ll measure
Build time and costs; visitors and acquisition sources; meaningful actions and return visits; revenue and ongoing expenses. Each published result will have a measurement window and a source.
Before the clock starts
- Validate the audience and the comparison problem.
- Define a small, useful first version.
- Connect measurement and verify it works.
- Launch publicly and record the start date.
This brief describes the planned public experiment. It does not report launch results or certify the status of any separate product build.
The process
Six steps. No skipping the first.
AI helps us build faster.
This helps us make better bets.
- 01
VALIDATE
Talk to a specific audience. Study what they already use. Test why they would switch before committing to a large build.
Find a problem worth solving.
- 02
BUILD
Use AI to build a focused product. Record the time, tools, costs, and tradeoffs along the way.
Make the smallest useful thing.
- 03
LAUNCH
Ship a usable version with measurement in place. A public URL is the starting line.
Put it in front of people.
- 04
TRAFFIC
Try a small number of distribution channels. Measure sources and useful visits instead of guessing.
Test how people find it.
- 05
USERS
Look for repeat use, conversations, and real problems solved. A visitor is not automatically a user.
See who comes back.
- 06
REVENUE
Publish revenue alongside costs. Treat willingness to pay as a hypothesis to test early, then measure actual results.
Test whether value gets paid for.
A learning loop, not a promise of revenue. Test willingness to pay before you build, too.