I pivoted my mindset, along with the product.

I failed continuously for the whole decade building 20+ products.

Here's what I am doing now to make it right this time.

Everyone knows, when building a product these are mandatory things to keep in mind: (but forgotten) 1. You need to have a paying customer on day 1 2. You need to build B2B and not B2C 3. You need to build something that don't scale

But, as an engineer whose day job is to think about scaling from day 1, I have struggled a lot trying to think the other way. It's hard right? It's a habit that needs to be broken.

I used to build a product that can scale for 100k visitors, million visitors, etc., But ended up with just 1 user, which is just me.

The rules are as simple as the above, it's just needs to be written down and engrained in your mind and soul, not to do anything that does not fall under those 3 categories. If you don't you will end up losing a decade like me.

"Ten years have got behind you. No one told you when to run, you missed the starting gun." - Time (Pink Floyd)

Things are not rigid though, because sometimes you need to build a tech heavy product, that requires scrapping, doing some ML to process it, before giving it to the customer with a neat interface. Not at all. That's how those huge companies work. You should not.

Yes, you might feel like your product is tech heavy, but it doesn't have to be all that on day 1.

Let me give you an example: I started building MevinX early March, I had a great vision, that I want to allow people to read the content only that they want to read, avoid all the doom scrolling.

To grow on twitter, one has to stop doom scrolling, instead, they should be focusing on doing those 20+ replies, and 3+ posts every single day consistently, that what everyone says on X "That's how I grew my X account in 3 months to 10k followers", and that's true too, I don't deny that.

But, what I did is that, I wrote a beautiful scrapper to get tweets from your own account, use LLMs to figure out the tweets that are more relevant to you, and ignore the rest. I did 3 iterations on scrapper to ensure that my customers don't get blocked when they use my product, and I did 8 algo iteration to make sure the feed is relevant, and still failed. Yeah, still failed. It's not about the tech or anything, it's beyond tech.

What went wrong? Everything, literally everything. I spent 4 months, iterating the algorithm that did not work, but the crawler is efficient.

What's the point of all those 4 months except learnings? $0 revenue, and launch was pathetic with just couple of users who gave it a try and did not use it afterwards.

I thought keeping in touch with potential customers was good enough to make them use it later, but people whom I was in touch didn't care after the launch, because the product wasn't so great externally, but beautiful inside. It's like having a beautiful heart, but everyone looking at your suit. Nobody cares about your heart on first look.

How would I do it, if I have to do it again? This is where my mindset change happened. I need to make few initial customers happy, in a most preached way "Do things that don't scale", Connect with few people who are facing the above problems, give them a curated list of tweet that will help them engage and make better impressions, and gain more followers.

It's hard and it's not scalable at all, but you are now talking to a potential customer who is ready to pay for it, if it works well. I would iterate based on their feedback and make the next round of curated tweets better, because I now have figured out their taste in tweets, because AI isn't going to understand that on day 1 without us feeding the necessary information.

Build a simple classifier, feed a data, get the tweets link and share it with the users. This is a slight improvement over a period of time, I would still manually verify if the output make sense before sharing it. Iterate and improve.

This will help me convert those early users into customers who are ready to pay, because they see the growth in their followers (qualitatively, and not just quantitatively). And they will also see how their life has become simple on X.

What am I doing next? I don't want to pursue MevinX with this new approach, because I found that it may not work well, so I paused all the development, converted that into a script that can fetch my own tweet that are trending in the past 2 hours and send me those links into my telegram and I engage with those whenever I find something interesting. (will make it open source soon)

I pivoted Mevin into a different product, and I am already having talks with potential customers who are willing to pay. No, I am still not building a product yet, I am just giving them something to get started and build the product over time. They are already willing to migrate to my platform, which is already a win.

What you should be doing? Avoid all the pitfalls, and strictly follow the simple rules.

Want to learn more how I am building this new product in public? with my own personal experience of all the ups-and-down? If yes, hit follow, I am happy to share my decade worth of knowledge here, so you don't make those mistakes.