Ask HN: What did you learn from your recent project failure?

[raydenvm]

Everyone’s had a project that failed —startup, side project, or ignored feature. What was your recent failure, and what's your take away? Not just "validate early," but some insights that changed your approach. Curious to hear and learn from your experiences.

[whatamidoingyo]

That payment providers label certain projects as "high-risk"*, and will refuse to allow you to use their platform. For example, supplements. Stripe and PayPal will close your account. You're left with providers like CCBill, which require a $1,450 fee to setup Visa and Mastercard.

So, always read the terms of service of platforms you intend to integrate into your projects. And do it before you spend weeks or months developing the project :').

* - High risk involves dating, supplements, CBD, and various other things.

[chistev]

What makes them high risk

[whatamidoingyo]

I think this article will give you a better idea: https://stripe.com/resources/more/high-risk-merchant-account...

This one is worth checking out as well: https://stripe.com/en-mx/legal/restricted-businesses

I missed this the first time around, but it looks like they've also banned "Get Rich Quick Schemes":

"Get rich quick" schemes, including investment opportunities or other services that promise high rewards to mislead consumers; schemes that claim to offer high rewards for very little effort or up-front work; and sites that promise fast and easy money

Interesting.

[chistev]

Why don't they approve it? Because of potential scam?

[whatamidoingyo]

Yeah, pretty much. But they indeed approve some companies. I guess if you have a clean, well-established business, it won't be a problem. If you're a fresh startup, you're in trouble.

The only options left are paying the $1,450+ fee, or accepting crypto currency payments only (which I think most people won't bother going through the trouble of this to purchase creatine, lol).

[ben_w]

That there is no correlation at all between (mobile app) code quality and business success.

Not an anti-correlation: I'm not saying bad code is a good business; I'm saying it's no correlation, that it doesn't make much difference either way.

The worst code you can imagine and then some? Business awards and acolades in one case, business went under in another. Latest design pattern, CI, code review? Too slow for market in one case, strong market position in another.

https://benwheatley.github.io/blog/2025/02/26-14.04.07.html

[raydenvm]

At some point, I also learned that code quality does not really matter. Product quality has an impact, but not code quality. With AI, it will likely matter even less because code quantity dramatically increases at the expense of quality.

[cpach]

I can see how this can be true for commercial products. Would you say it also applies to open source projects?

[raydenvm]

You mean non-commercial open source, right? Well, perhaps... I can't say with certainty because I was never involved in one.

[yamirghofran]

Use dev containers and try to deploy a toy system (all the components but toy code) early on in the project. We did a lot of work and couldn't deploy Qdrant, even though it was working on localhost.

[raydenvm]

How much time/money would you save then?

[apineda]

even small projects cost a lot of time ( marketing !! ), better to focus on something you really care about rather than opportunistic items

[account-5]

I learned I hate docker and elastic stack.

[eyesofgod]

To just not bother with side projects. Got too costly too quickly.

[raydenvm]

Have you ever felt tempted to drop your main job because your side project was growing?

[aristofun]

1. There is no link between quality and popularity of the product. As long as the product is just good enough - it is only a function of marketing.

2. There is no universal lessons to learn, or wisdom to apply in every situation. Including the point above. Every complex enough situation is unique. “bad generals prepare for the past war” as someone said.

[raydenvm]

1. What about the concept of minimum loveable products (MLP) versus minimum viable products (MVP)? They say go for MLP and that with MVP, your chances are much lower due to the excessive product and information noise.

[ben_w]

I say that everyone doing a startup needs to know what the actual minimum in the "minimum viable product" really is for them.

The idea of the MVP is there to say "publish this as soon as you can, don't wait because the market will have moved on and you'll run out of money if you wait for perfection".

MLP is a reaction against this, but I don't know if it's a reaction by perfectionists (who are wrong because then you go bankrupt), or a reaction by people who've tried one too many buggy and unready alpha products that was shipped to test the market before it was ready (who are correct, I'll let you think of your own example here).

[aristofun]

It’s just bunch of word fluff for people to show/think they are smart. In reality money or/and luck usually wins.