Ask HN: Tools and techniques to quickly validate startup ideas | Hacker News Reader

Ask HN: Tools and techniques to quickly validate startup ideas

4 pointsthrowawyfunder3 comments

Do any of you have any recommendations for tools and techniques to quickly validate startup ideas? Especially the ones that you don't have connections for. I am interested in AI ideas that could potentially help other industries but I have no means to validate how painful the problem is. Some of the approaches that I am considering:

- Post a job posting to recruit some employees and gather information from them (and pay them for their time) - Create a landing page and advertise it to the right target audience - Reach out to people through LinkedIn

Any other tools, techniques, recommendations, etc., or even suggestions on how to make effective use of these strategies?

(throwaway because my main account is easily identifiable to work colleagues)

I'd be very careful about implimenting AI in a field you're wholly unfamiliar with. It sounds like you've got a solution that's looking for a problem, so at the very least I'm glad you're trying to validate it. On the other side of the coin though, there is a massive chance you could miss your mark, or worse yet, exacerbate issues that are already present in AI and (further) threaten the credibility of AI-based products.

Good call. Yeah, I am worried about building something that either doesn't serve users well or worse cause other unintended problems. As a technical person, getting to know new technologies seems a lot easier than getting to know new industries and markets. And that's why I am looking for ways to understand industries better.

Make a list of risky assumptions. Sort by impact on being wrong about them. Now you have a list of the riskiest assumptions.

Test those assumptions with experiments (google ads, landing pages, etc) that provide data. Set a bar for success. Run experiment.

If results > success, move on to the next riskiest assumption. If results < success, form a new hypothesis, rinse, and repeat.

Reply on news.ycombinator.com