How do you test an idea before talking to users?

I keep it scrappy. I just run quick tests with friends who share the niche. It is not perfect research but it shows me reactions early and saves time before setting up formal interviews.

Hi Elena! I've done deep dives on reddit as you said, and actually gathered a fair bit of info from social media using keyword searches. I found X and TikTok very lucrative for information, as well as facebook and linkedin (although my product is very corporate orientated).

Hope this helps!

You actually can use an LLM for this, though the answers will generally skew positive if you're asking questions like, "Do you like this?" Better would be to try a prompt like, "I have this idea, how would [define your target audience] use this?" I like using multiple LLMs to see where the consensus seems to be across the answers given, rather than rely on just one.

Do you have any resource for an audience that would be interested in the idea? If so, I've also seen people generate a very basic site with a "sign up to learn more when this is launched" and a brief description to see if people sign up as a gauge of interest. I think the trick is to have an audience you've already built, though, and for that audience to have the target market you're aiming for with your idea - otherwise, this is extremely difficult to do.

Depending on the idea and the market you're trying to reach with it, you could consider the website "sign up if you want to learn more when it launches" idea and going to local businesses in your area to see if they have bulletin boards where you could post a one-pager with a QR code for people to sign up, as well. Super old school, but better than investing in a set of interviews.

I've also tried Reddit searches and scrapes of posts to find pain points that my offering would address, versus testing my idea outright. Sometimes, it's in the positioning rather than the idea itself where you'll gain traction.

1. Test it on my own after some time (so I had enough time to forget what happened with the product and see some mistakes clearly)

2. Ask a relative to test it (because he/she is gonna have the first experience) – just observe

3. Reach out to ICP (3 – 5) – have a call with them and observe what they do

4. Roasting on social media (good luck) 😀 – esp. on Reddit

5. Paying for qualified tester

6. Ship it publicly and receive feedback via forms/chat etc.

Before interviewing people, kill the riskiest assumptions fast:

1. Run a smoke test (landing page + waitlist or “fake door” CTA), make sure to have analytics in place.

2. Then mine Reddit or X reviews to map pains and language for copy and positioning yourself.

3. Simultaneously, do a quick “Mom Test” on 3-5 ICPs. Ask about past behavior and workflows, not opinions, to avoid any false sugarcoated reviews before deeper calls.

If it passes those two gates, you can do some concept or usability tests with lightweight tasks to spot show‑stoppers early on. I did this on a bulk Video Editing tool for creators, and it worked magically.

I'm launching another app next month, btw (failed subtle plug)

I would say a little bit from all areas :)

I think you would start with a market analysi, look for trends and for predictions for upcoming years. But most importantly, you should start with the problem!

I usually start by talking to a handful of potential users who are approachable and willing to share honest feedback. My goal isn’t to “sell” the idea, but to verify my assumptions and see if the problem I’m addressing is real for them.

A helpful framework I often refer to is The Mom Test — it’s a great book about how to talk to users without leading them or asking questions that just validate your bias.

For me, it’s often as simple as writing down a few problem statements and pressure-testing them with quick polls or posts in relevant communities. If people raise their hand or start sharing their own experiences, that’s usually enough validation to know the idea’s worth taking into real interviews.

I’ve usually started with two things before jumping into interviews:

1. Reverse-engineering reviews/Reddit threads to map out frustrations and language people already use.

2. Sketching a lightweight business plan (even just in bullet points) to pressure-test assumptions (basically asking myself “does this logic hold if I strip it down?”)

That second part is actually why we’re building Escape Velocity AI: it helps founders stress-test their plans and assumptions before they spend time/money chasing the wrong thing. Right now, we’re keeping it free while we gather feedback, so if you’re interested in testing approaches like this, we would love to hear your take.