My GPT, Ranked Top 360 Among 3 Million GPTs Worldwide, Shares Some Insights
A tech guru compiled a global Top 500 GPT ranking based on usage data, and my earliest GPT, "GPT Finder," ranks 360th. Considering there are over 3 million GPTs in the GPT Store now, my GPT is truly a rare gem.
However, in reality, it only took me about half an hour to create it. In this article, I will detail the journey of creating this GPT, revealing the key to making a successful and widely-used GPT. Additionally, since my billing address is in the U.S., OpenAI mentioned in their blog that they will pay GPTs Builders in the U.S. based on usage starting in Q1 2024. If you're curious about how much the Top 500 GPTs can earn, stay tuned.
Experience One: Create Something You Truly Need
In the field of internet products, there is often a conflicting debate about whether a product manager should be a typical user of their own product. Some argue that it should be the case, as only by truly using your own product can you fully understand what users need; others contend that it shouldn't be, because you can only represent a small segment of users, and being overly focused on your own perspective might lead to overlooking the needs of actual users. Both viewpoints are valid, but if you are new to product management or just starting out with independent development, your best bet is to become a user of your own product. Compared to guessing what others need, which often requires strong empathy and user understanding skills, it is much easier to understand yourself. You are more aware of the characteristics and boundaries of the need, especially if the product you create is something you truly need and find useful. Since you are usually not that unique, this need often resonates with a large number of people. When I had the idea for this GPT, the official GPT Store had not yet been launched, and numerous navigation sites had emerged, which I neither wanted to nor had the time to browse one by one. Moreover, many GPT navigation sites at the time did not even have search functions, so creating a GPT that could search for other GPTs was a perfectly reasonable choice.
Experience 2: Launch Quickly, Build MVP and Iterate
MVP stands for "Minimum Viable Product." Given how easy it is to create with GPT, and considering that many needs are likely to be thought of by more than just me, you should do everything you can to get your product out there faster. For example, with my GPT Finder, there are actually many ways to search for GPTs. For instance, I could write a crawler to scrape all GPTs and build a database, then integrate it with my GPT through Actions' API. However, writing a crawler, building a database, and writing an API would be quite challenging to accomplish quickly. Another approach would be to organize GPTs data into a table and upload it as a knowledge base to GPT. The issue here is that GPTs were being added so rapidly that constantly updating the table would be a significant task. Therefore, I adopted a rather lazy approach, allowing my GPT to search for other GPTs based on certain rules. The specific instructions for this method were mentioned in my introductory video on GPTs. You can view it through this link: https://www.bilibili.com/video/BV1jC4y1E73E/
Experience Three: Build in Public
It means publicly building your product on social media. For example, after completing this GPTs, I posted it on Jike, shared it on WeChat Moments, and also created related video introductions on YouTube and Bilibili. This public building process can promote my product, gain initial seed users, and understand user needs to obtain necessary feedback for further iterations.
By the way, because of this public building process, this GPT also received a recommendation from a YouTube influencer with 140,000 followers, Jojo, which undoubtedly brought a lot of traffic to my GPT.
If you currently lack a fan base to enable this public building process, you are also welcome to self-recommend the GPTs you created in the comments below the video. It is expected to gain some attention. Moreover, if I find it genuinely good, I will also create a dedicated video to introduce everyone's GPTs.
In fact, in addition to this GPT Finder, my AI assistant has been used by over 900 people. The "Mangge Master" for conversing with Charlie Munger and the "Language Bridge" for professional English-to-Chinese translation have also received over 400 and 300 uses, respectively.
You may have seen many bloggers analyze the GPT Store and create tutorials for making GPTs, but the GPTs they make themselves often don't yield much result. After all, developers and content creators belong to two groups with very different characteristics. I dare to say that I am one of the few bloggers who truly practices and shares sincerely with you.
If you are interested in learning truly practical AI usage experiences, welcome to subscribe to my "ChatGPT Mastery Guide" column via: https://xiaobot.net/p/AIclass and join my AI learning community. GPTs created by my community members will have a better chance of being recommended by me.
Appendix: Below are a few GPTs I've created that I'm quite satisfied with:
- GPTs Finder: Helps you find the GPTs you need: GPTs Finder
- Master Munger: Engage in conversation with Charlie Munger: Master Munger
- Language Bridge: Professional English to Chinese translation assistant: Language Bridge
- Stable Diffusion Prompt Generator: Generates prompts for Stable Diffusion: Stable Diffusion Prompt Generator
- Amazon SEO Expert: Essential for cross-border e-commerce: Amazon SEO Expert
- Minimalist Logo Design: Designs logos and explains the design philosophy: Minimalist Logo Design
- AI Reading Assistant: Quickly understand books and get summaries and reading assistance: AI Reading Assistant