The real leak was after signup, before the first useful action
An tool for pulling clean web content for LLMs and agents got 331 signups over about 7 weeks, around 47 per week. Of those users, 322 created an API key. Only 112 ran their first scrape, which is the the product exists to support.
Then 34 started a trial, and 23 became . Of the , 20 converted from a trial and 3 subscribed directly. The biggest leak happened between creating an API key and running the first scrape.
After signup, users landed on a , had to copy the API key, switch to a or , send a real request, and read the result. The product had a playground, but it was one click away instead of being the first thing users saw.
Key points
- 331 people signed up over about 7 weeks.
- 322 created an API key, but only 112 ran the first scrape.
- 34 started a trial, and 23 became .
- The main came after setup but before the first real use.
- The playground existed, but it was not placed directly in front of new users.