How do you know when you're wrong?

The reality is, you'll be wrong often so the best thing to do is (1) make educated guesses using data or "secret" you've uncovered and (2) test those ideas as quickly as possible. Testing assumptions doesn't always mean building a product. Sometimes it's as simple as talking to potential users or running a quick wizard of oz test. Of course, all startup advice is contextual and the way to approach a particular problem varies widely.

Here is a quick stab at the question from creating a startup/product perspective. TL;DR: Check out the 5 Why approach (https://en.wikipedia.org/wiki/5_...) to identifying the root cause of a problem and building a product that solves a problem. This is also similar to the first principle problem-solving approach shared by Elon Musk (https://jamesclear.com/first-pri...). Using these approaches, you can increase your odds of being correct when people think you are wrong, unfortunately, you probably can never be fully confident you are correct. Finally, for a startup to be successful, getting the product right is a necessary but not sufficient condition. You need to get many other things right as well (e.g., go-to-market strategy) LONG VERSION Typically, entrepreneurs listen to a problem and jump to building a product/solution. If you are able to break the problem into principle components and solve the root cause, even if people around you think you are wrong, you "might" be right (emphasis on "might") However, if you are at the top level of the problem, and jump to solving a problem, and people think you are wrong, there is a very high probability you are wrong. Let's take an example of the 5 why approach. (BTW - this is a simple made-up example, but hopefully gets the point across.) - Problem: It costs a lot to create a video highlight reel Why? - A Video Editor (high hourly billing rate) has to spend significant time to create highlight reel (First Why) Why? - A Video Editor needs to watch the full video multiple times before creating the highlight reel? (Second Why) Why? - It is necessary to watch the full video to identify the interesting parts of the video? (Third Why) Why? - Video Editor uses judgment and context to identify interesting parts while watching the full video? (Fourth Why) Why? - Judgment and context are important to identify interesting parts of the video? (Fifth Why) Potential solution: Automate highlight reel clip prediction (judgment) for specific video types using artificial intelligence and machine learning.

I don't think it's good to think in terms of wrong or right. It's too polarized and will cause a cascading directional loop. What you think is either wrong or right is just a moment in time, and compounding on that isn't a good think. The road is too windy during product development. I believe it's far better to march towards your vision, right or wrong, until you find a fit. A fit is simply more right than it is wrong. In that sense, I think continuous iteration is king. It's like playing Mario Kart with blinders on. If you're right, it's right. If you're wrong, you at least know it's not right.

You: thought about it for a really long time, from seemingly all angles, but fewer angles than you realize at the time with what felt like little epiphanies along the way that kept you fired up to stick with it. Other People: thought about it for 10 seconds and have an opinion, you recognize this and think, obviously I thought of that too after 10 seconds, you're useless. Then some data. Then you realize you're not even close to having figured it out completely, and that you could have gotten a lot more out of other person's "instant" response/thought. With enough grinding and iterative work, "right and new and special and worthwhile" breakthroughs emerge.

Terms like product market fit and pivot probably come to mind when reading your post. Probably plenty of articles on these matters because it depends on a lot of factors (team size, funding, etc). I'm still pre-launch and my conviction hasn't waivered in 5 years.

I think it's just about finding a problem that's non-obvious. I'm thinking 40-60% problem validation potentially means you're on to something. Too little validation and it probably isn't a big enough problem and too much validation probably means there are already alternative solutions. Ben Horowitz says you need to have a good idea that at first glance looks like a bad idea.

I find my best nudge to rethink things is when I have a strong emotional response to the information I'm receiving. If a user gives me feedback and my instinct is to become angry, reject the feedback, or rationalize it away, I know there's a real problem. In the absence of real users, I benefit from giving myself sample scenarios to walk through with the product and emphasizing each point of interaction or observation. This helps avoid the pitfalls of breezing through the interactions that I've probably already tested hundreds of times. I liken this to reading a piece of writing out loud to find the mistakes, rather than skimming.

When people don't like what I make. But that doesn't mean the problem is wrong...just that I might need to iterate and adjust the solution.