You know the saying, "The first impression is the last impression."
For someone using your product or open source project and learning it the very first time, is it any good?
How do you know? Are you measuring it?
How do you measure what "good" is?
I would start with identifying what the minimum viable desired outcome is from your audience and measure that. Call it Time to Minimum Viable Desired Outcome (TMVDO – or "tim video"). Ideally, you'd identify this by interviewing and getting feedback from developers.
Here's an example...
For a game engine and a new-to-gamedev developer, it's most likely "time-to-first-game." What does that entail? Many steps, from installation to conceptual explanations to practical code snippets. This could be in the form of a tutorial, mini-course, video series, or full lab environment.
Once you've designed that experience, in order to improve it over time, you have to measure its performance. Not code performance but learning performance. Not only, "How long does it take?" but: where do people get stuck? Drop-off? Become frustrated? Go back?
If you want to create a good first impression, you'll have to define what that means and measure the living f*ck out of it. If you don't, you'll be relying on fuzzy page analytics or the gracious souls who take the time to tell you it doesn't work in a GitHub issue.