“Strategic Moves Catapulted This GPT To The Top Of OpenAI’s Charts!”
As you probably already know, OpenAI’s GPT Store launched today: https://chat.openai.com/gpts/
The store lists 8 categories as a leaderboard:
1. Top Picks
2. DALL·E
3. Writing
4. Productivity
5. Research & Analysis
6. Programming
7. Education
8. Lifestyle
My custom GPT called “SEO” (https://chat.openai.com/g/g-GrshPDvS3-seo) currently ranks as #7 in “Research & Analysis” — see main image.
And this wasn’t by accident — I had a plan!
After listing to OpenAIs Dev Day (https://devday.openai.com/) on 6 November 2023, it became clear that OpenAI was going to concentrate on custom GPTs (what they referred to at the time as Assistants, which are now the API version of GPTs)
Based on that, I decided to convert my 16 existing ChatGPT Plugins (https://twitter.com/OrrenHimself/status/1724752345693856142/) into custom GPTs (mentioned: https://twitter.com/OrrenHimself/status/1721713616221794651) as well as learning how to build GPTS (https://orren.medium.com/what-everyone-else-isnt-telling-you-an-insider-s-guide-openais-custom-gpt-assistants-8656c9cd8244).
Then around the same time, a few custom GPT Directories popped up to help people find custom GPTs as there was no search functionality within ChatGPT to facilitate this.
“SEO” only had 110 uses 7-days from custom GPTs launching (https://twitter.com/OrrenHimself/status/1724409635267076125)
I knew that the discovery of custom GPT for those who didn’t have an audience (especially an audience that 1) already had ChatGPT Plus, 2) used custom GPTs within ChatGPT Plus and 3) used the niche GPTs I created) seemed to be coming from directories (https://twitter.com/OrrenHimself/status/1724296666151600517) or existing user bases.
If you look at the GPT Store now (link previously) you will see “Ai PDF, ScholarAI, Canva, Humanize AI, Web Pilot, AskYourPDF, Consensus, KeyMate.AI. Mixerbox etc were all existing Plugins from the Plugin Store that were converted to GPTs by their creators just like I did.
This likely means they either had 1) loyal users (like my “SEO” one, or user databases of an audience that 1) already had ChatGPT Plus, 2) used custom GPTs within ChatGPT Plus and 3) used the niche GPTs they created because they required OAuth 2 logins (https://orren.medium.com/solved-oauth-for-chatgpt-plugins-so-you-can-charge-for-usage-47b98d28828c) to use.
I only had the former — loyal users.
I also knew that OpenAI would replicate the search function at some point with Custom GPTs (https://twitter.com/OrrenHimself/status/1724331720772862020), just as they did with Plugins — recall with plugins there was no search feature when it launched, but later added one (https://twitter.com/OrrenHimself/status/1724296671050489862)
As I had developed 16 Plugins before OpenAI stopped taking submissions (https://twitter.com/OrrenHimself/status/1732613570599518237), I knew the custom GPT Store, just like the Plugin Store would reward custom GPTs that have the most uses by placing them in the featured part of the store (https://twitter.com/OrrenHimself/status/1724296668911407192).
And, as we know from other app stores, it’s a self-fulfilling cycle — those who are in the top categories tend to stay in the top categories because people assume they are useful because they are in the top categories.
So based on this, my plan was to submit my GPTs to as many directories (I ended up collating a list of 18) as possible, as soon as possible to boost their uses in case this is how the custom GPT Store would work.
And when it launched today, that is how it works!
And my top GPT by uses, “SEO”, was the one that featured in the Top 12 in “Research & Analysis”.
Interestingly, the number of uses OpenAI showed me it had privately via https://chat.openai.com/gpts/mine before the GPT Store launching was accurate (I was tracking uses via a web app through Actions because I run ads on them.)
Those accurate stats I had privately, then got reset on the day of the launch.
You can see how many it has by following this methodology (https://twitter.com/OrrenHimself/status/1737816931003244818/).
The number you can see publicly right now is less than what it had before the store launched and got it into the store
So, no 1000+ is not the minimum number of uses to get in the Top 12, it’s much higher than that (don’t ask me what it is because I won’t tell you, sorry.)
I offered (here: https://twitter.com/OrrenHimself/status/1724336968686071988 and many other places) the list of 18 directories to anyone who wanted it and to copy my strategy.
Very few took me up on the offer.
